x

Mobile Skype: Quality Issues?

VOIP peer-to-peer pioneer Skype on Tuesday made a big move by porting its popular desktop VOIP client to certain wireless handhelds. Sounds neat, but now that the news has digested, the big question appears to be: Will it really be worth it?

Skype's introduced a PDA version that will work with 802.11b-enabled Pocket PC handhelds with more than 40 Mbytes of memory onboard (see Skype Me? Skype You! on the regular Skype client).

In theory, this is an application that enables users to -- at least partially -- bypass wireless carriers: If users live in an area with plentiful WiFi hotspot access and all their friends and family use Skype.

That's a big if, if you consider the nature of hotspot coverage when compared with traditional mobile networks: It's spotty, at best...

How spotty? Get all the details on Unstrung.

— Dan Jones, Site Editor, Unstrung

dljvjbsl 12/5/2012 | 1:55:51 AM
re: Mobile Skype: Quality Issues? Yep. data propellerheads? You mean the ones that created SONET? CO switching? Unix? DWDM?

Wow, and you stormed the beaches at Normandy, too? Anything else you'd like to take credit for


Technonerd, he was just mocking you.

The technologies that he mentioned all came from the telephone industry. They came along with the dedication to service, quality, customers that was taught to me when I joined the industry those many years ago.
technonerd 12/5/2012 | 1:55:56 AM
re: Mobile Skype: Quality Issues? Yep. data propellerheads? You mean the ones that created SONET? CO switching? Unix? DWDM?
Wow, and you stormed the beaches at Normandy, too? Anything else you'd like to take credit for?
mavericktel 12/5/2012 | 1:55:59 AM
re: Mobile Skype: Quality Issues? dljvjbsl,
Yeah he's probably a troll with too much time on his hand. A Dilbert. He doesn't even understand the basics of transmissions yet he acts like he's some old school Bellhead. He's like your weird Uncle who says things and no one bothers to challenge him because he just barks and eventually you realize he's loony.
mavericktel 12/5/2012 | 1:55:59 AM
re: Mobile Skype: Quality Issues? That's an obvious point technonerd doesn't understand. If you don't even understand TCP/IP and why UDP is the only choice rather than any connection-oriented protocol since it has lower "overheard" in terms of getting your voice encapsulated packets from one endpoint to another, then trying to enter the debate is moot.
My understanding is that UDP is best for packet video conferencing because dropped packets aren't retransmitted. I can hire people to figure out the rest of it.

Hire people. LOL. You sort of got it right.


TDM is simply Time Division Multiplexing. PCM is Pulse Code Modulation. Then we have to discuss sampling ( 2x the frequency minimum per Nyquist theorem), companding, and then we arrive at mu or a-law for basically relying on a log-type scaling to capture voice samples. This is all part of circuit switching.
We don't have to discuss any of these things. Unless, by chance, we are data propellerheads wishing to hide behind tech-talk to avoid the real issues at hand. Gotta admit, you guys are good at that.

Why not? Can't even understand Old School voice?

The PSTN is built upon TDM with its DS hierarchy.Even SONET rings which CO's sit on as nodes rely on SONET's (US version) basis on the DS3 formatas the basic payload for its payload envelope.

That's why gateways are required to hand off voice traffic to the PSTN. These boxes have to accept let's say PRI voice T1 circuits from the Telco and they also rely on codecs to do TDM/VoIP conversion.
O.K., now what was the point of going through this? Because I used the "map to" insead of "convert?" Geez, no wonder data propellerheads are such dweebs.

Yep. data propellerheads? You mean the ones that created SONET? CO switching? Unix? DWDM?
lastmile 12/5/2012 | 1:56:04 AM
re: Mobile Skype: Quality Issues? Mobile quality issues are really not of great concern to Skype at this time. Skype mobility is an offshoot of the initial success of the original P2P Skype. The quality is astounding and there is a steady increase in the number of users. I have noticed figures as high as 350,000 at any one time.
350,000 is by no means a small number. The number of users are far in excess of that figure.
Let us not discuss quality issues for stuff that is free.


aswath 12/5/2012 | 1:56:15 AM
re: Mobile Skype: Quality Issues? So the total bandwidth upstream is reasonable, unless there are an awful lot of subscribers/node.

The other liming factor is the backhaul bandwidth available from CMTS. I read somewhere that in South korea the ratio between backhaul interface and the access interface is very high compared to US. Can anybody provide any empirical data? (Apparently it was stated in WTF 2004 hosted by Isenberg.)
fgoldstein 12/5/2012 | 1:56:19 AM
re: Mobile Skype: Quality Issues? 348> The upstreams on the MSO DOCSIS plant are fairly small pipes and you can't do many simultaneous VoIP sessions using G.711 uLaw codec technology.

Well, not really a problem in typical use. While a DOCSIS 1.1 upstream is limited in bandwidth (typically set to 5 Mbps, often 2.5M), it's common for a CMTS to have four upstream carriers per downstream. So the total bandwidth upstream is reasonable, unless there are an awful lot of subscribers/node. (Okay, we do know of some such cable operators.) DOCSIS 2, of course, provides for much more upstream bandwidth, provided the modems are also at 2.0+.
dljvjbsl 12/5/2012 | 1:56:22 AM
re: Mobile Skype: Quality Issues?
O.K., now what was the point of going through this? Because I used the "map to" insead of "convert?" Geez, no wonder data propellerheads are such dweebs.


I promised myself that I would not point out any of your howlers again but this is too juch. Have you nay idea about transmission and signalling? You have not given any indication that you know anything about encoding or switching. Indeed you have given every indication that you do not.

Tell us teh truth. is this just one big troll?
stephenpcooke 12/5/2012 | 1:56:44 AM
re: Mobile Skype: Quality Issues? mavericktel wrote:

The PSTN is built upon TDM with its DS hierarchy. Even SONET rings which CO's sit on as nodes rely on SONET's (US version) basis on the DS3 format as the basic payload for its payload envelope.

That's why gateways are required to hand off voice traffic to the PSTN. These boxes have to accept let's say PRI voice T1 circuits from the Telco and they also rely on codecs to do TDM/VoIP conversion.


SONET is my backyard. DS3 is A supported payload but so is anything that is 'bulk-mapped' providing that you have a mapper card to do the mapping/de-mapping for you, it can be literally anything. ATM has its own standardized mapping into an STS-1/STS-3c/STS-12c/..., DS-1 has its own, Packet Over SONET (POS) has its own, etc.

Yes, the PSTN is based on T-circuits and SONET transport but neither of these limit the choice of higher layer traffic other than maximum provisioned bandwidth, which can be increased by adding more equipment/wavelengths, etc. as necessary. Most newer SONET equipment has cards that do all sorts of mapping/de-mapping and it is quite likely that some equipment providers will leverage their installed base to offer additional capabilites for quite some time. The only protocol conversions necessary are due to the class 5 switch, not the transport mechanisms.
technonerd 12/5/2012 | 1:56:44 AM
re: Mobile Skype: Quality Issues? Some technology will succeed. ATM to the desktop? Ferggetaboutit.
ALERT! Religious debate among engineers dead ahead!


It's just the 'Net and applications are not as simple and closed off as the PSTN so potential security exploits are always a problem. Years ago, a very bright fellow who was an old school data comm veteran mentioned that VoIP will open up a can of worms in terms of security issues.
Just wait until these enterprises that have moved to VoIP have their phones killed by some 16-year-old hacker; the "consultant" who sold it to them is nowhere to be found; and the IT staff wants to shut the whole thing down for the rest of the day.


Video is not the killer app. It's VoIP.
The sole reason for VoIP is regulatory arbitrage of one sort or another.


If consumers want silly cameras stuck in their phones, so be it.
You mean you're not doing up-skirt photography with your cellphone? What kind of engineer are you? Don't you realize you could be locked out of the next IEEE meeting?!
technonerd 12/5/2012 | 1:56:44 AM
re: Mobile Skype: Quality Issues? That's an obvious point technonerd doesn't understand. If you don't even understand TCP/IP and why UDP is the only choice rather than any connection-oriented protocol since it has lower "overheard" in terms of getting your voice encapsulated packets from one endpoint to another, then trying to enter the debate is moot.
My understanding is that UDP is best for packet video conferencing because dropped packets aren't retransmitted. I can hire people to figure out the rest of it.


TDM is simply Time Division Multiplexing. PCM is Pulse Code Modulation. Then we have to discuss sampling ( 2x the frequency minimum per Nyquist theorem), companding, and then we arrive at mu or a-law for basically relying on a log-type scaling to capture voice samples. This is all part of circuit switching.
We don't have to discuss any of these things. Unless, by chance, we are data propellerheads wishing to hide behind tech-talk to avoid the real issues at hand. Gotta admit, you guys are good at that.


The PSTN is built upon TDM with its DS hierarchy.Even SONET rings which CO's sit on as nodes rely on SONET's (US version) basis on the DS3 formatas the basic payload for its payload envelope.

That's why gateways are required to hand off voice traffic to the PSTN. These boxes have to accept let's say PRI voice T1 circuits from the Telco and they also rely on codecs to do TDM/VoIP conversion.

O.K., now what was the point of going through this? Because I used the "map to" insead of "convert?" Geez, no wonder data propellerheads are such dweebs.
technoboy 12/5/2012 | 1:56:52 AM
re: Mobile Skype: Quality Issues? Re Post 358:

The statement that VOIP designers dont recommend G.711 is not true. Every VOIP vendor that I have seen and tested with has g.711 as an option. The scenario is that users typically use G.711 for LAN/MAN applications and then use either G.711 or G.729 for traffic that traverses a WAN. The G.722 codec will begin to replace G.711 as new IP phones are sent out to the market. This is considered a wideband codec and is used for IP endpoint to IP endpoint conversations.
mavericktel 12/5/2012 | 1:56:52 AM
re: Mobile Skype: Quality Issues? Video is not the killer app. It's VoIP. I'll repeat it again. I don't care much for VoIP solution even as a replacement for TDM-based PBXs but cell phones and their lack of 5 "9"s
have mass appeal. Cellular's claim to fame is simply mobility. Why would anyone want to have broadband access through their cell phone is beyond me. It's not whether or not I like a piece of technology but I do believe VoIP is the killer app just as e-mail once was for the 'Net and now search engines. If decision makers of businesses choose VoIP solutions then so be it.
If consumers want silly cameras stuck in their phones, so be it.
technoboy 12/5/2012 | 1:56:52 AM
re: Mobile Skype: Quality Issues? RE Post 355

My reply after checking out the URL is that I dont think this will be a necessary technology going forward. There are several development efforts that are designed to create a dual modem technology that would allow users to roam on 802.11 networks and handoff to a mobile network once they leave their campus or building etc... (SIP) is a key piece of this capability since it will play in all aspects of the network. I think having a mobil router in your car for voice only is probably unecessary ( of course cisco would probably disagree). That does not mean a similar device may not be useful if it is serving multiple functions.
mavericktel 12/5/2012 | 1:56:53 AM
re: Mobile Skype: Quality Issues? "Packet's the way to go, but only if they do it right."

They already have. I have seen businesses use different ISPs due to geographic distance so they chose local ISPs. I didn't realize it was basically using the Internet was the WAN until I inquired about the ISDN bandwidth. It was during the day time.
mavericktel 12/5/2012 | 1:56:54 AM
re: Mobile Skype: Quality Issues? ISPs deal with DOS attacks all the time on their customers. They have to backtrack and filter the traffic. Phreakers took advantage of the PSTN's inherent weak points. It's just the 'Net and applications are not as simple and closed off as the PSTN so potential security exploits are always a problem. Years ago, a very bright fellow who was an old school data comm veteran mentioned that VoIP will open up a can of worms in terms of security issues.
mavericktel 12/5/2012 | 1:56:55 AM
re: Mobile Skype: Quality Issues? There's WiMax and iBurst. IP won out though it has many flaws. So did Ethernet over Token ring.
Some technology will succeed. ATM to the desktop?
Ferggetaboutit.
mavericktel 12/5/2012 | 1:56:55 AM
re: Mobile Skype: Quality Issues? IP has a completely different purpose than PCM. The distinction you are trying to make is between TDM and packet (IP).

That's an obvious point technonerd doesn't understand. If you don't even understand TCP/IP and why UDP is the only choice rather than any connection-oriented protocol since it has lower "overheard" in terms of getting your voice encapsulated packets from one endpoint to another, then trying to enter the debate is moot.
Don't need to rely on any guru to figure it out
if you understand data communication.

TDM is simply Time Division Multiplexing. PCM is Pulse Code Modulation. Then we have to discuss sampling ( 2x the frequency minimum per Nyquist theorem), companding, and then we arrive at mu or a-law for basically relying on a log-type scaling to capture voice samples. This is all part of circuit switching.

"Maps PCM to IP?" What sort of lame pseudo data prophead speak am I hearing you utter? lol.

The PSTN is built upon TDM with its DS hierarchy.
Even SONET rings which CO's sit on as nodes rely on SONET's (US version) basis on the DS3 format
as the basic payload for its payload envelope.

That's why gateways are required to hand off voice traffic to the PSTN. These boxes have to accept let's say PRI voice T1 circuits from the Telco and they also rely on codecs to do TDM/VoIP conversion.
mavericktel 12/5/2012 | 1:56:56 AM
re: Mobile Skype: Quality Issues? The upstreams on the MSO DOCSIS plant are fairly small pipes and you can't do many simultaneous VoIP sessions using G.711 uLaw codec technology.

G.711 if I recall is toll-grade. No one in VoIP design recommends that codec. I forgot which Cisco or Nortel recommends. All the same codecs they have to choose to implement.
technonerd 12/5/2012 | 1:56:56 AM
re: Mobile Skype: Quality Issues? Google was another search engine but it is now the defacto leader.
Google was new. They figured out what legions of data geeks never bothered to realize because they were too busy cleaning the ear wax out of their ears with their paperclips: Those boolean operators were unnecessary. A simple insight that made all the difference.
technonerd 12/5/2012 | 1:56:58 AM
re: Mobile Skype: Quality Issues? Who cares about pictures? VoIP is here
Who cares about VoIP? Voice has been here for 100 years. If that's all the propellerheads can come up with after all this time and all this funding, it's pretty damn sad.


Nortel has been pushing their Succession VoIP/TDM solution.
They've been doing that for nearly a decade now. Who knows, maybe someday they'll get it right.


Most traditional PBX vendors have been working on VoIP. CTI
Oh God, not CTI!! Talk about beating your dead horses. No one wants it except for call centers. It has about as much appeal to the mass market as on-line grocery shopping and "the intelligent house." Which is to say zero.
mavericktel 12/5/2012 | 1:56:58 AM
re: Mobile Skype: Quality Issues? Ever been on a video conference on an international level utilizing BRI's?
Yes, in fact I have on more than one occasion. It wasn't BRIs, it was always a PRI line on each side. The quality was mediocre.

-Doesn't matter if it's PRI or BRI since most vidconf. CPE only typically expects certain bandwidth anyway or ISDN bearer channels.

Telephone is good enough, Instant messaging, and e-mail. Do I need holographic projections of you in 3D for communication?
The history of telecom is that when you send more information the functionality rises. Voice is the equivalent of a command-line interface. Video is the equivalent of a GUI. Holography? Let's get pictures right first.

-Who cares about pictures? VoIP is here. Video will be perfected in terms of quality where end-users find them acceptable beyond the stuttering ISDN video frames. I've seen one school trying out off the shelf to videotape courses to videostream over the'Net. They ran it on their local LAN for testing and it was quite excellent.
I would of never seriously considered digital cameras stuck in cell phones to be of any use but go figure, people use it.

Sheesh, you're forced to use UDP, because IP is ever present and is the de facto data protocol suite of choice. You can create new protocols.

From what I've been told (and had demonstrated), there's no need to create new protocols. UDP works great.

Then what is RTP? UDP is simply convenient.
It encapsulates the voice payload.

Problem is that it's not new, so the VCs and investment banks would have a harder time selling it as revolutionary

You don't need to. Google was another search engine but it is now the defacto leader. Microsoft tried to buy them out but now will build their own search enginer into their longhorn OS. You've totally missed the point.
Nortel has been pushing their Succession VoIP/TDM solution. Most traditional PBX vendors have been working on VoIP. CTI (ex. Dialogic)
were already working on PC integration with voice handoff from the PSTN (not VoIP however).

Cisco AVVID ring a bell?
Packet Man 12/5/2012 | 1:57:08 AM
re: Mobile Skype: Quality Issues? Technonerd in Msg#27 wrote:
Problem #4 is also a voice killer. Wi-Fi doesn't allow for handoffs, which means it can only be portable rather than mobile. What does someone do with their Wi-Fi phone in their car? Nothing. It won't work in their car, because the car is moving and you can't do handoffs. Not only that, but if they ever did develop handoff capability could you just imagine the difficulty of managing the number of handoffs involved in a typical mobile session given the small sizes of cells?

My comment:
While not really meant for WiFi TODAY, it will only be a matter of time. See: http://www.cisco.com/en/US/pro...

Some day this will be miniturized and will easily fit and work inside of a WiFI or other wireless device....and VoIP calls will not get interrupted... its only a matter of time.

PM
Packet Man 12/5/2012 | 1:57:08 AM
re: Mobile Skype: Quality Issues? In response to MSg#11....

Wheter it be wired or wireless.... free Internet access in all hotel rooms, is getting very popular up here in Canada. Many hotels are offering this now, and most are anticpated to in a few years.
PO 12/5/2012 | 1:58:16 AM
re: Mobile Skype: Quality Issues? TRA have some decent courses, and they tend to keep their technical details "technically" accurate. (Nobody's perfect.)

But their courses are developed by individuals with their own views on things, and thus don't tend to have a comprehensive view of what's out there.

So ... while they often provide a decent introduction, I would not recommend using them as an authoritative source.
technonerd 12/5/2012 | 1:58:28 AM
re: Mobile Skype: Quality Issues? What's CIF? One might accuse you of being a propeller-head acronym slinger. *grin*

Wow, you really know how to hurt a shit missilier, don't you? :-)

From the Telecommunications Research Associates "Lexicat":

The standard video conference signal format used in the ITU-T H.261 standard. Allowing different video codecs to communicate among each other using the common format. The format is 288 lines/frame and 360 pels/line for the brightness signal.

By the way, a commercial for Telecommunications Research Associates. They're not my "guru" on anything, but they are the best source I know of for training in solid, basic telecom infrastructure concepts.
priam 12/5/2012 | 1:58:37 AM
re: Mobile Skype: Quality Issues? Right. It would have been more exact to day eBGP. iBGP is commonly multihopped, but being in your own domain, should be something you don't have to worry about wrt DOS, - anyway, it takes less trouble to prevent it. eBGP is as you say external, and is usually to routers one hop away. There are various reasons to multihop, but even there, - though it opens a wider hole - you can filter TTL < 255-N.

->I believe the reason the vast bulk of BGP sessions are a hop away is that customers purchase Internet circuit from different ISPs and require BGP for redundancy purposes to handle inbound traffic in case of one circuit failure. So most of the BGP sessions are run between customers as a stub network off the edge of ISPs. BGP sessions can be controlled by network engineers who simply allow certain configurations to be implemented on their edge routers connecting to customer routers.

eBGP is run AS to AS (AS is Autonomous System) and iBGP is run internally.
dljvjbsl 12/5/2012 | 1:58:37 AM
re: Mobile Skype: Quality Issues? Do you know that VoIP can and does carry voice encoded in PCM. Do you know that that the PSTN, just like VoIP, can and does carry it in other encoding schemes as well. To call the PSTN digital network, the PCM network, as you do, shows that you really do not.

technonerd replies
It's late at night so I can't remember the specific standard, but yes I know of the one that maps PCM to IP

How else would you think that a packet network would carry voice than in a voice coding standard. PCM, ADPCM and other standrds which I do not know much of anything about are used to carry voice and analog data across the network. In fact in a VoIP call, there may be multiple codecs used each for their own advantages.

IP has a completely different purpose than PCM. The distinction you are trying to make is between TDM and packet (IP).

dljvjbsl 12/5/2012 | 1:58:38 AM
re: Mobile Skype: Quality Issues?
Compression codecs on the PSTN have pretty much gone away as bandwidth has gotten cheaper. It's been a pretty long time since I've seen ADPCM anywhere. Even voice mail systems stopped using compression codecs now that gigantic disk drives cost next to nothing. Wireless is the space where you see compression codecs. You'll also be seeing them on the MSO VoIP systems once they go to mass deployment. The upstreams on the MSO DOCSIS plant are fairly small pipes and you can't do many simultaneous VoIP sessions using G.711 uLaw codec technology.


Technonerd -- do you see now that multiple codecs are used and that thye are chosen for a variety of reasons. ADPCM (adaptive delta PCM) is a common codec and also note that it is a form of PCM
alchemy 12/5/2012 | 1:58:38 AM
re: Mobile Skype: Quality Issues? dljvjbsl writes:
Do you know that VoIP can and does carry voice encoded in PCM. Do you know that that the PSTN, just like VoIP, can and does carry it in other encoding schemes as well.

Compression codecs on the PSTN have pretty much gone away as bandwidth has gotten cheaper. It's been a pretty long time since I've seen ADPCM anywhere. Even voice mail systems stopped using compression codecs now that gigantic disk drives cost next to nothing. Wireless is the space where you see compression codecs. You'll also be seeing them on the MSO VoIP systems once they go to mass deployment. The upstreams on the MSO DOCSIS plant are fairly small pipes and you can't do many simultaneous VoIP sessions using G.711 uLaw codec technology.
alchemy 12/5/2012 | 1:58:39 AM
re: Mobile Skype: Quality Issues? technonerd writes:
And 384 Kb/s for ISDN through the tandem (which I suspect is how most military video conferences go, although I may well be wrong on that) is still limited to CIF, which might have been o.k. at one time but is looking pretty old nowawdays.

What's CIF? One might accuse you of being a propeller-head acronym slinger. *grin*

The problem with Switched 384 service (other than being stinkin' expensive as heck) was that it only worked on the AT&T network. If you didn't have a PRI line to a #4ESS, you couldn't get the service. This is all moot since the #4ESS is now obsolete and AT&T is in the process of moving everybody off those switches unless a customer is using a service that isn't supported by their other interconnect solutions.
alchemy 12/5/2012 | 1:58:39 AM
re: Mobile Skype: Quality Issues? technonerd writes:
The history of telecom is that when you send more information the functionality rises. Voice is the equivalent of a command-line interface. Video is the equivalent of a GUI. Holography? Let's get pictures right first.

I pretty much agree with this. I think the real issue is the price of the devices and access to big pipes with good QoS at a low price.

In the early days of commercial videoconferencing (PictureTel, Compression Labs; H.320; nx64 with BONDING or switched 384) both the access device and the telecom pipe were very expensive. Video Conferencing Units were $10K to $50K, you had a hefty fixed charge for your BRI or PRI lines, and you paid a stiff per-minute charge for the bandwidth. Video Conferencing was only for the big corporations and, even there, the bean counters weren't too happy about using it too much.

Today, the pricing model is quite different. I have a VideoPhone sitting on my desk at the office. It cost $600 and the price is dropping very fast. The VideoPhone has a crappy little screen that's fairly useless but it also has an output port to pump the video image into a monitor. Using a 1/2 megabit/sec codec speed and an external monitor, I get very good image quality on the corporate network. At 7AM when the corporate network connection to the internet isn't being used and the internet is fairly quiet, I get excellent image quality to remote locations.

The missing piece of the puzzle is dynamic QoS across the internet. The technology exists but it won't get deployed unless people are willing to pay for it. I don't think VoIP is going to be enabler since people will stick to traditional PSTN networks if they're going to pay for the service. Video may very well be the service that people are willing to pay for that will allow the technology to be deployed.
technonerd 12/5/2012 | 1:58:40 AM
re: Mobile Skype: Quality Issues? Do you know that VoIP can and does carry voice encoded in PCM. Do you know that that the PSTN, just like VoIP, can and does carry it in other encoding schemes as well. To call the PSTN digital network, the PCM network, as you do, shows that you really do not.
It's late at night so I can't remember the specific standard, but yes I know of the one that maps PCM to IP. And yes, I know that at least the core of the PSTN will take other encoding schemes. When I used "PSTN" and "PCM" interchangeably, I refer to the PSTN's main purpose, which is to carry PCM voice traffic from one end to the other.

As for our relative contributions, dljvjbsl, I think you're oh-so-typical of the arrogant data propellerheads who designed solutions than don't solve, features that don't work, equipment that doesn't equip and applications that don't apply. And then, should anyone have the temerity to note that the new is worse than the old, you resort to what dataheads always resort to: A stream of acronyms and tech-talk to cover up the underlying failure of the system to do even a small fraction of what you promised it would.
technonerd 12/5/2012 | 1:58:40 AM
re: Mobile Skype: Quality Issues? Anybody who was serious about video conferencing connected directly to an AT&T #4ESS on Primary Rate and ran switched 384 (H0 or 6-channel) calls. That solution was bullet-proof.
A VERY expensive way to go, which is why IP is far preferable. And 384 Kb/s for ISDN through the tandem (which I suspect is how most military video conferences go, although I may well be wrong on that) is still limited to CIF, which might have been o.k. at one time but is looking pretty old nowawdays. Packet's the way to go, but only if they do it right.
dljvjbsl 12/5/2012 | 1:58:40 AM
re: Mobile Skype: Quality Issues?
Yes, I have heard of everything you've mentioned, and no, I'm not going to go into a dissertation on them here because it's pointless.


You have previously demosntrated that you know very little if anything about private networks. It is apparent from your answer that you know even less about trasmission and voice quality. This even though you rail about the lack of quality on IP networks. There is, of course, nothing wrong with that. I come to these boards to learn from the knowledgeable and often do research after visiting here to learn about important issues.

Do you know that VoIP can and does carry voice encoded in PCM. Do you know that that the PSTN, just like VoIP, can and does carry it in other encoding schemes as well. To call the PSTN digital network, the PCM network, as you do, shows that you really do not.

Why don't you listen, ask questions and contribute what you do know to the discussions here. The constant launching of missiles as you do is pointless.
alchemy 12/5/2012 | 1:58:41 AM
re: Mobile Skype: Quality Issues? technonerd writes:
I was also told that, with ISDN-based video conferencing units, there has been a frequent problem with channels dropping in and out. I am told that the units are hard to set up and hard to use, and of course that the connection charges are very high.

For a variety of reasons, I was effectively persuaded that the PSTN is not the way to go with video conferencing and never will be on any sort of mass-scale basis.


The channel dropouts resulted from slips in individual channels. In BONDING mode 0, you calculate the relative delay between channels at call setup time and have no way of detecting if the relative delay changes during the call. In (I'm a little rusty... this is 10 year old technology) BONDING mode 2 and, maybe mode 1, the data is striped with an overhead pattern and you can recalculate relative delay on the fly and add/drop channels. Most implementations only used mode 0.

Anybody who was serious about video conferencing connected directly to an AT&T #4ESS on Primary Rate and ran switched 384 (H0 or 6-channel) calls. That solution was bullet-proof.

nx64 ISDN calls were a non-starter because a #5ESS couldn't do the feature without sticking a hack (and expensive) switching matrix in front of the #5ESS that could do the feature.
technonerd 12/5/2012 | 1:58:43 AM
re: Mobile Skype: Quality Issues? Ever been on a video conference on an international level utilizing BRI's?
Yes, in fact I have on more than one occasion. It wasn't BRIs, it was always a PRI line on each side. The quality was mediocre.


Telephone is good enough, Instant messaging, and e-mail. Do I need holographic projections of you in 3D for communication?
The history of telecom is that when you send more information the functionality rises. Voice is the equivalent of a command-line interface. Video is the equivalent of a GUI. Holography? Let's get pictures right first.


Sheesh, you're forced to use UDP, because IP is ever present and is the de facto data protocol suite of choice. You can create new protocols.
From what I've been told (and had demonstrated), there's no need to create new protocols. UDP works great. Problem is that it's not new, so the VCs and investment banks would have a harder time selling it as revolutionary.
technonerd 12/5/2012 | 1:58:43 AM
re: Mobile Skype: Quality Issues? Non-delay? Technonerd, what do you think video is? It's even worst since VoIP simply requires good enough objective/subjective rating of voice quality and VIDEO requires BOTH voice & video (the visual aspect - "stuttering" frames of video etc). That's where the beauty of CODECs comes in. Talk to your gurus about this. lol.

Maybe you're an engineer, because only an engineer could be that incoherent.
mavericktel 12/5/2012 | 1:58:50 AM
re: Mobile Skype: Quality Issues? Maybe they were wrong, but it's a lot more likely that I simply misunderstood them. In any case, I think it's pretty well known that ISDN video conferencing has not been successful. It's hard to use; the reliability of the connections isn't high enough; the quality of the pictures and synchronization with voices isn't good enough.
--

You're absolutely incorrect. Ever been on a video conference on an international level utilizing BRI's? Even Internet access utilizing different ISPs provide relatively high-quality
videoconf.


I'm convinced that PSTN video conferencing cannot be fixed for a cost that would permit its widespread adoption in the mass market. I am convinced that, for video conferencing to be viable, it must be done via packet. I typically have written "IP," but actually the people I trust tell me UDP is the way to go because there's no retransmission of lost packets in UDP.
---

What are you talking about? It's the expensive videoconf systems that are hard to set up. Most end users require system integrators to handle the set up. Why would I need videoconferencing?
Telephone is good enough, Instant messaging, and e-mail. Do I need holographic projections of you in 3D for communication?

No, go back and first learn what TCP/IP as a suite IS and what types of packets/datagrams are USED for. Ask your buddies what RTP is. UDP is connection-less. IP is as well. Connection-oriented transmission require SYN and ACKs.
I send you something, you respond in a given amount of time. You don't respond, I believe you didn't receive and I retransmit that packet of chunk of data. This affects latency and jitter though this can be mitigated by buffers. All of this depending on how the data traverses what type of network ( may require reorder of packets sent, etc.)

Sheesh, you're forced to use UDP, because IP is ever present and is the de facto data protocol suite of choice. You can create new protocols.
One carrier runs IS-IS instead of iBGP typical of Carrier's IP backbone.
mavericktel 12/5/2012 | 1:58:50 AM
re: Mobile Skype: Quality Issues? technonerd,
hate to burst your bubble. But CU See Me videoconf app over the internet was created
ages ago. It was freeware and since most
user were on dial-up, it was useless for
me but if you had a Univ lab w/ high speed Internet access, it wasn't bad.

You have to handle latency & jitter with buffers anyway. I have some books/materials with algorithms to calculate proper VoIP on LAN/WAN connections and networks.

Non-delay? Technonerd, what do you think video is? It's even worst since VoIP simply requires good enough objective/subjective rating of voice quality and VIDEO requires BOTH voice & video (the visual aspect - "stuttering" frames of video etc). That's where the beauty of CODECs comes in. Talk to your gurus about this. lol.
mavericktel 12/5/2012 | 1:58:50 AM
re: Mobile Skype: Quality Issues? technonerd,
Do you have any conception of how technology works? You know like how software is written?
Why do you think Microsoft is always allowing
you updates for security patches? Yep, potential
exploits technonerds/hackers ( the original term)
figured out and relayed to MS.

Hey I can walk over to your cube at work and if you don't leave your workstation "locked" if you have win2000 workstation, yeah, I can look at your files and maybe attach some interesting file and email it to my email box. Sheesh. lol.

Sort of like an auto maker allowing any key to open any of their cars. With software, the more sophisticated, the harder it is to properly model how a piece of software will behave under any given condition.

What you should be more concerned about is spyware. It tracks your surfing activities and sends it back to a home server. If it can assign that to your email, name, any personal info, more the better. One security expert checked and found clear text messages of personal info sent by spyware. Even your ISP may give your spyware
when you install the generic app to connect online.
mavericktel 12/5/2012 | 1:58:50 AM
re: Mobile Skype: Quality Issues? Also, a year or so ago someone observed that the vast bulk of BGP sessions are between routers within a hop or two. That gave rise to an operational tweak: accept no TCP segment on a BGP connection with IP TTL < ~224. This makes it very hard for an attacker from the edge to disturb a connection
---

I believe the reason the vast bulk of BGP sessions are a hop away is that customers purchase Internet circuit from different ISPs and require BGP for redundancy purposes to handle inbound traffic in case of one circuit failure. So most of the BGP sessions are run between customers as a stub network off the edge of ISPs. BGP sessions can be controlled by network engineers who simply allow certain configurations to be implemented on their edge routers connecting to customer routers.

eBGP is run AS to AS (AS is Autonomous System) and iBGP is run internally.
technonerd 12/5/2012 | 1:58:58 AM
re: Mobile Skype: Quality Issues? You talk about delay that is undetectable to the ear. Have you ever heard of the phrases echo and ERL and LNE quality measures? The reason that the delays in these connections are impercepible to the ear, is that there has been a great deal of research and some exceedingly good engineering and maintenance to make it so

So THIS is the underlying reason for your white-flame outburst before? All I can say is you labored mightily before giving birth to a mouse. Yes, I have heard of everything you've mentioned, and no, I'm not going to go into a dissertation on them here because it's pointless.

Your point is ambiguous so maybe I'm misreading, but it appears that you took my prior post as an attack on the PSTN for having any delays. If so, that was not what I intended to convey. Delay is a reality, and the PSTN does a great job of dealing with it. But the people I trust on the video conferencing side of things have told me that PSTN delay is a significant problem for ISDN-based systems.

Maybe they were wrong, but it's a lot more likely that I simply misunderstood them. In any case, I think it's pretty well known that ISDN video conferencing has not been successful. It's hard to use; the reliability of the connections isn't high enough; the quality of the pictures and synchronization with voices isn't good enough.

I'm convinced that PSTN video conferencing cannot be fixed for a cost that would permit its widespread adoption in the mass market. I am convinced that, for video conferencing to be viable, it must be done via packet. I typically have written "IP," but actually the people I trust tell me UDP is the way to go because there's no retransmission of lost packets in UDP.
technonerd 12/5/2012 | 1:58:58 AM
re: Mobile Skype: Quality Issues? First off, I think (unusually, for this thread) that technonerd was overstating the problem of differential delay in ISDN. This was known from the get-go; ISDN standards support signaling for n*64 channels with 8 kHz integrity (no differential delay) but that is rarely implemented. Instead, the Bandwidth ON Demand INteroperatbility Group created BONDING, a hardware-oriented standard for delay equalization. Its main use was indeed video conferencing.

It was implemented in a bunch of PRI-type and multi-BRI devices. The data world (dial up ISP use) mostly used MLPPP instead. The ISDN newsgroup was really full of confustion in those days when people said "bond" their B channels together, or "bonding" them, when they meant MLPPP, not BONDING.


Thanks for the explanation. You may well be right. We all have some subjects that we know better than others, and frankly when it comes to differential delay I rely heavily on what others have told me. I could imagine having a misunderstanding.

I was also told that, with ISDN-based video conferencing units, there has been a frequent problem with channels dropping in and out. I am told that the units are hard to set up and hard to use, and of course that the connection charges are very high.

For a variety of reasons, I was effectively persuaded that the PSTN is not the way to go with video conferencing and never will be on any sort of mass-scale basis.
whyiswhy 12/5/2012 | 1:59:11 AM
re: Mobile Skype: Quality Issues? Look, this exact issue (vulnerability of the national com infrastructure) is exactly why we need some government oversight and regulation. Otherwise it is very likely that in the rush to provide the cheapest and most easily provisioned bandwidth, one vendor (with his latest wiz bang idea) makes it possible to bring down the whole system.

What we get instead is a hands-off, leave it to the market attitude. The government builds GIGBE, and leaves the rest of the country to suffer its fate.

Cheap, not smart.

I mean, to put a new toilet in your house takes plans and permits and inspectors. And that system was put there for public health reasons, and to keep consumers from getting ripped off.

For one, I consider my telephone (and xDSL line) much more important than my toilet. And potentiallly much more dangerous.

Where is the common sense? When we had a regulated monopoly, it was easy...the monopoly took care of those problems.

The FCC should be all over this like white on rice. I mean, national security and all.

Just venting.

-Why
dljvjbsl 12/5/2012 | 1:59:13 AM
re: Mobile Skype: Quality Issues? I would agree with everything you said about SS7 with the caveat that it is not impossible to mount DDOS attacks on it.

As said before, radio station contests can trigger events like this on the PSTN. The PSTN attempts to protect itself from this eventuality both adminstratively (special numbers for radio stations etc) and dynamically (speacial COS for emergency numbers etc.).

However if we are positing an attack be an army of zombie PCs, in my opinion anyway, the attack on the PSTN network would be much easier and more effective. Suppose they all dial a speciic number at the same time.

I know of one case that I spoke of before where a man mounted an attack on the 800 number of an evangelist he thought was exploiting his mother.

For reasons, that I do not understand this attack was difficult to trace. I would ahve thought that the ANI information would have provided identification immediately. However the 800 number charges from the incessant hang up calls was reported as significant which leaves out the lost productivity of the call center.
fgoldstein 12/5/2012 | 1:59:14 AM
re: Mobile Skype: Quality Issues? Two topics.

First off, I think (unusually, for this thread) that technonerd was overstating the problem of differential delay in ISDN. This was known from the get-go; ISDN standards support signaling for n*64 channels with 8 kHz integrity (no differential delay) but that is rarely implemented. Instead, the Bandwidth ON Demand INteroperatbility Group created BONDING, a hardware-oriented standard for delay equalization. Its main use was indeed video conferencing. It was implemented in a bunch of PRI-type and multi-BRI devices. The data world (dial up ISP use) mostly used MLPPP instead. The ISDN newsgroup was really full of confustion in those days when people said "bond" their B channels together, or "bonding" them, when they meant MLPPP, not BONDING.

Security: I think a DDoS attack can knock out any server on the Internet, such as a SIP proxy, just from sheer volume of crud, just getting there. There are defenses, but that raises the stakes again. The PSTN's SS7 network has no 100 Mbps Ethernet ports going to customers. It's mainly 56 kbps and 1.5 Mbps links, very tighly controlled.

Topologically, the Internet has a fair number of dumbbells -- wide at the edges and narrow in the middle. It also has footballs -- wide in the middle, narrow at the edges. Footballs congest at the egress, dumbbells at the ingress. SS7 networks have fewer large speed transitions. It's mainly a football, with no wide access points. Thus DDoS is essentially impossible.

There are good reasons to sometimes run SS7 over IP, but the core of the network remains blissfully isolated. It is technically competitive, with multiple providers (Internet-like), but membership is rather exclusive.
technoboy 12/5/2012 | 1:59:15 AM
re: Mobile Skype: Quality Issues? Post 326 reply

Great Post!!!
technoboy 12/5/2012 | 1:59:16 AM
re: Mobile Skype: Quality Issues? Stephencooke

So, already the doors are wide open on the PSTN. All the discussion about 'trusted' domains and proxies is moot. Cryptographically, they are wide open because the encryption is not even turned on!!! Aren't we humans interesting beasts?

Can you elaborate on this statement? You were refering to the WI-FI networks I believe but are you trying to compare that to something else as well?
technoboy 12/5/2012 | 1:59:16 AM
re: Mobile Skype: Quality Issues? Post 319:Do you? I thought the largest of the multiple delays will be the delay. Will IP network offer smaller delay? Don't you have to absorb jitter by introducing additional delay?
This is not the way it was explained to me, anyhow. Maybe my guru got it wrong, or maybe I heard it wrong, but I was told that you add them up. And then you add a bit of time for reassembly.

In a VOIP Network you have an end to end delay budget that you engineer to. The delay budget consists of cumulative impact of delay associated with the following IP endpoint delay ( measured in milliseconds)+delay on the LAN( typically measure in microseconds)+the serialization delay associated with each router hop in the network (measured in milliseconds)+ delay associated with the actual distance between any two endpoints ( measured in microseconds). This would be for a call between two IP endpoints. Within the endpoints the delay has to do with converting voice into an IP format and it is dependent upong the codec selection and sampling time. Jitter buffers in IP phone can be dynamic and are usually set for 30 milliseconds although some can be set for longer times to account for the delay associated with extremely long distances between endpoints. The total cumulative impact of all of this delay should not exceed 150 milliseconds one way in order to achieve good voice quality. Service provider networks that are delivering VOIP are engineered to support end to end delay of less than 90 milliseconds. This is done with a combination of removing the notion of hop counts from the core of the network by using MPLS and moving BGP to the edge of their networks. Also QOS and security mechanisms are also done at the edge of the network rather than on a hop by hop basis. We recently tested a new IP phone on a service providers network between philadelphia and belgium. The phone used a G.722 codec which is considered a wideband codec even though it still transmists at 64Kbps plus IP overhead. With this setup we demonstrated the difference between a call at G.711 and G.722 with QOS tured on and off. The end to end delay never exceeded 120 milliseconds and the variance was between 107 and 120 milliseconds which was well within the 30 milliseconds jitter buffer.
stephenpcooke 12/5/2012 | 1:59:17 AM
re: Mobile Skype: Quality Issues? This seems to imply that security is taken much more seriously by the research community and the vendors than by actual users.

Isn't this pretty much what several of us have been saying from the start? This is probably the most common 'engineering' mistake in the book. We design all sorts of great things assuming that people will somehow be told how to use them and will use every feature that we work so hard to produce.

When I got married, I worked for Nortel who had great deals on their phones for employees. There were several that had the features that I was looking for so I made the mistake of calling my wife-to-be to get her view on things. Her only question was "What colour is it?" We got the most basic phones in the colours that made her happy.

So, already the doors are wide open on the PSTN. All the discussion about 'trusted' domains and proxies is moot. Cryptographically, they are wide open because the encryption is not even turned on!!! Aren't we humans interesting beasts?

The initial quote sounds like a perfect setup for technonerd's favorite VoIP rant:
Stupid idiot newbie users! What do they expect, something that works like we said it would? How naive could they be? What -- they think they're just going to punch a few buttons and talk, fer chrissakes?
priam 12/5/2012 | 1:59:17 AM
re: Mobile Skype: Quality Issues? The integrity of BGP peering sessions has been actively considered for quite some time now. RFC 2385, which defines MD5 authentication for BGP, was a standard in 1998. (BTW, it's authentication, not encryption, and it does happen at the TCP layer, via adding a new TCP option, which permits adding a MD5 digest over a TCP segment, including data and header.)

Also, a year or so ago someone observed that the vast bulk of BGP sessions are between routers within a hop or two. That gave rise to an operational tweak: accept no TCP segment on a BGP connection with IP TTL < ~224. This makes it very hard for an attacker from the edge to disturb a connection.

Actually, given reasonable filtering at any network boundary, an attacker ought not to even be able to reach a BGP session.... but there you go.

->The questions that will be asked are how quickly was the workaround arrived at? How complete a fix is it? Is this just another kluge or has it been thought through so we won't have another firestorm on our hands next week? Given the relative isolation of the PSTN there was time to make the changes, in the case of IP it is less likely that this will be the case. In the meantime how much damage will have been done to corporate reputations?
dljvjbsl 12/5/2012 | 1:59:18 AM
re: Mobile Skype: Quality Issues? Well, you've said I was wrong. Now show it.

You talk about delay that is undetectable to the ear. Have you ever heard of the phrases echo and ERL and LNE quality measures? The reason that the delays in these connections are impercepible to the ear, is that there has been a great deal of research and some exceedingly good engineering and maintenance to make it so.

I certainly have heard lines and trunks which had poor performance in these areas both in the lab for subjective tests and in the field. Delay is a highly important parameter in the estimation of quality.

I can assure you that both the PSTN and the VoIP networks take this issue very seriously and have remedial measusres to cope with it.
dljvjbsl 12/5/2012 | 1:59:18 AM
re: Mobile Skype: Quality Issues?
please provide even one example of a deployed system where VoIP control signalling is encrypted. At least in the PacketCable world, the MSOs have all turned 3DES off since they can't debug the network otherwise.


This seems to imply that security is taken much more seriously by the research community and the vendors than by actual users.

I rad an artile in the Daily Telegraph (from London) about the experience of a columnist who used the WiFi service supplied by McDonald's at a restauurant in Mnahattan. He found it easy to log on and the service useful. He even found that it was a complimentry service at least for a while.

With all of this he noted that he really did not need the McDonald's service. His WiFi modem detected 24 open networks in the area. He could have used any of them

One wonders that if these companies do not seem to care and are suffering no ill effects then where is the concern about security that everyone talks about.

Countering this was the test performed at the IETF meeting in San Diego. I do not know the details but the WiFi system there was monitored to see which types of exploits would be used against it. From a presentation there and at other places by Steve Bellovin, the system was under constant attack by known exploits from the time it was turned on.

This is a curious situation.
alchemy 12/5/2012 | 1:59:19 AM
re: Mobile Skype: Quality Issues? stephenpcooke writes:
VoIP control signalling is encrypted and in-band. How long before a hacker figures it out?

Err... please provide even one example of a deployed system where VoIP control signalling is encrypted. At least in the PacketCable world, the MSOs have all turned 3DES off since they can't debug the network otherwise. I haven't even seen an MSO who has turned on DOCSIS BPI+ encryption on the access network.
technonerd 12/5/2012 | 1:59:24 AM
re: Mobile Skype: Quality Issues? Do you know anything about telephony or at least transmission? I have never seen so many howlers in one place.

Channel 4 and channel 5 not being synchronized in a synchronous network -- you don't even know the issues. I know that we are all here to learn from people who know more. But for someone who has insulted just about everyone about IP for reasons of transmission quality (QoS) and you do not even know the rudiments of transmission delay, this is just amazing.

I have never seen so many howlers in one place and you were the one who posted the definition of PCM to educate us.


Well, you've said I was wrong. Now show it.
dljvjbsl 12/5/2012 | 1:59:27 AM
re: Mobile Skype: Quality Issues?
What really bugs me on the video front is that we don't have this stuff widespread already. The technology is readily available, but the group thinkers just don't see it. That's why I think some foreigners without Silicon Valley's blinders and preconceptions will get there first.


You can buy this stuff (delay binding) from just about any company that has voice processing products. I was designing this stuff in the late 70s and early 80s and I was not young then.
dljvjbsl 12/5/2012 | 1:59:29 AM
re: Mobile Skype: Quality Issues?
will try this one more time. Yes, the PSTN is great as QoS within any given channel. But it cannot synchronize between channels. You have "channel 2" doing its thing well; you have "channel 4" doing its thing well; you have "channel 5" doing its thing well. But channel 2's timing isn't synchronized to channels 4 and 5 and so on.

The delays -- which are present, but simply undetectable to the ear -- are not coordinated. So the equipment often has to wait before it can reassemble. The result: lip-synch problems, picture problems, etc.


Do you know anything about telephony or at least transmission? I have never seen so many howlers in one place.

Channel 4 and channel 5 not being synchronized in a synchronous network -- you don't even know the issues. I know that we are all here to learn from people who know more. But for someone who has insulted just about everyone about IP for reasons of transmission quality (QoS) and you do not even know the rudiments of transmission delay, this is just amazing.

I have never seen so many howlers in one place and you were the one who posted the definition of PCM to educate us.

Why don't you ask questions instead of launching missiles. You might learn something and this would be a more pleasant place.
technonerd 12/5/2012 | 1:59:29 AM
re: Mobile Skype: Quality Issues? Do you? I thought the largest of the multiple delays will be the delay. Will IP network offer smaller delay? Don't you have to absorb jitter by introducing additional delay?
This is not the way it was explained to me, anyhow. Maybe my guru got it wrong, or maybe I heard it wrong, but I was told that you add them up. And then you add a bit of time for reassembly.


On a lighter note, it is refreshing to see technonerd defending IP networking instead of PSTN.
Oh, I understand why you would feel this way given the impression I've created. (Oh the burdens of the shit-missilier ...) But in point of fact I've said many times that I think IP is useful for non-delay-sensitive traffic, and that IP video conferencing would (in my opinion) be a breakthrough product that would probably justify all the expense and lies associated with the voice-over-IP crapola.

What really bugs me on the video front is that we don't have this stuff widespread already. The technology is readily available, but the group thinkers just don't see it. That's why I think some foreigners without Silicon Valley's blinders and preconceptions will get there first.

Any bright young Indian engineers with rich parents reading this? Kid, the world lies at your feet for the taking. Really.
dljvjbsl 12/5/2012 | 1:59:30 AM
re: Mobile Skype: Quality Issues?
I've rarely seen a better description of how and why DDoS attacks succeed. Any individual source falls within acceptable and normal behavior. Collectively, their target is neutralized.


Can you be more specific. What level of calling is going to cause these problems? Switches are designed for appropriate behavior in overload state. They might not do it well but the proxies will have strategies to cope with mass calling.

I can see the SIP proxies being much smaller than the calss 5s. This to should mitigate this risk
aswath 12/5/2012 | 1:59:31 AM
re: Mobile Skype: Quality Issues? Msg. #312: When you mux them back at the machine on the other end, you have to wait for everything to arrive before you can reassemble it. Which means that you wind up adding all the delays together -- at best.

Do you? I thought the largest of the multiple delays will be the delay. Will IP network offer smaller delay? Don't you have to absorb jitter by introducing additional delay?

On a lighter note, it is refreshing to see technonerd defending IP networking instead of PSTN.
PO 12/5/2012 | 1:59:31 AM
re: Mobile Skype: Quality Issues? ... security in IP networks consists, figuratively speaking, of people running around and plugging holes in the dikes.

Before we all get too depressed, we go happily about our daily lives with this being true of most security. There are new defects found in cars and tires, but we still drive. There are break-ins around the city, but we still leave home. Security is always relative, never absolute. There are muggings, but we don't hide.

Security for VoIP has focused on 'average' behaviours, and less on whatever misbehaviours might do to the end-user's experience. It'll get there if it goes cautiously.
aswath 12/5/2012 | 1:59:32 AM
re: Mobile Skype: Quality Issues? From Msg #293:

These systems are notorious for QoS problems of various kinds, and those problems trace back to concantenation. The PSTN is great for individual 64 Kb/s streams, but don't dare rely on binding those channels together the treaing the result as a river.

I wonder whether you know this on your known or from the famed person in the know. I don't buy it. After all the PSTN gives a low, bounded jitter, with no loss. So if one can realize the original stream in an IP network, then most certainly it can be done using PSTN as well. If the current systems don't do it, then it may be because they are built for a circuit-switched H0 channel instead of thinking of 6 independent B-channels and running a multilink protocol on top of them.

Economics is another matter. For some reason, high speed symmetric DSL is a fraction of tariffed T1. But as you argue for voice, any difference in tariff can most likely be attributed to regulatory factors.
technonerd 12/5/2012 | 1:59:32 AM
re: Mobile Skype: Quality Issues? After all the PSTN gives a low, bounded jitter, with no loss. So if one can realize the original stream in an IP network, then most certainly it can be done using PSTN as well. If the current systems don't do it, then it may be because they are built for a circuit-switched H0 channel instead of thinking of 6 independent B-channels and running a multilink protocol on top of them.
I will try this one more time. Yes, the PSTN is great as QoS within any given channel. But it cannot synchronize between channels. You have "channel 2" doing its thing well; you have "channel 4" doing its thing well; you have "channel 5" doing its thing well. But channel 2's timing isn't synchronized to channels 4 and 5 and so on.

The delays -- which are present, but simply undetectable to the ear -- are not coordinated. So the equipment often has to wait before it can reassemble. The result: lip-synch problems, picture problems, etc.

From having seen and used the ISDN stuff and talked with people who will forget more than I'll ever know about it, I am convinced that if you're going to do video conferencing on any kind of scale it has to be IP video conferencing.


Economics is another matter. For some reason, high speed symmetric DSL is a fraction of tariffed T1. But as you argue for voice, any difference in tariff can most likely be attributed to regulatory factors.
This is because customers are willing to pay more for the QoS embodied in voice, not to mention the SS7 features. A data lines gives you best-efforts service to the best-efforts Internet. To the extent that QoS of any kind is introduced to the data network, the price goes up.
technonerd 12/5/2012 | 1:59:35 AM
re: Mobile Skype: Quality Issues? We need to look at security as a package, and never pretend that just putting a lock on the door solves the problem. VoIP technology is fine, but the protocols are not ready to be used across a subscriber interface in an all-out deployment

I couldn't possibly be less convinced than I am now that VoIP is ready for prime time fronm a security standpoint. There are new IP vulnerabilities discovered each and every day. I am totally convinced that, at the moment, security in IP networks consists, figuratively speaking, of people running around and plugging holes in the dikes.
technonerd 12/5/2012 | 1:59:36 AM
re: Mobile Skype: Quality Issues? Let's not blame the network for the failures of ISDN marketing, pricing, and adoption. The worst that can be said about ISDN is that engineers, not consumers, decided what the data rates should be. And that having to recondition lines that were longer than the network operators thought added to ISDN's cost burdens. But a lot was learned.
You are mixing apples and oranges here. The loop conditioning issues emanating from 2B1Q are interesting and I know a great deal about them. But they are irrelevant to this particular discussion. In fact, ISDN itself is only tangentially relevant.

I mention "ISDN" in connection with wideo conferencing equipment because, by virtue of the video stuff having been very expensive it was sold primarily to business users. Such users invariably have PBXs, and of course they are invariably connected with ISDN-PRIs.

The PSTN's problems with respect to videoconferencing are not failures of ISDN per se. The problem is that, when you demux a stream into four or six or eight or however many DS-0s you're going to use, each of those DS-0s takes a separate switching path.

This introduces timing differences. When you mux them back at the machine on the other end, you have to wait for everything to arrive before you can reassemble it. Which means that you wind up adding all the delays together -- at best. This, in simplified form, is why the PSTN generally sucks at video.

There are some super-broadband workarounds for it, but they are hideously expensive and complicated. The military uses them, but virtually no one else does for the same reason the military will pay $600 for a hammer while not even the biggest corporation will do so.


Similarly, a T1 circuit was always priced relative to the value of a DS0: the cost of channelized bandwidth was kept high because of its dual-use nature. Data services over DSL created a new pricing opportunity because the voice (PBX trunking) revenues were not (directly) threatened by DSL lines.
Excuse me, but "DSL," as you put it, refers simply to line codes used for Layer 1 transport. In today's market, when a DS-1 is delivered over copper that copper uses as "DSL" code in Layer 1. Whether the service is data to and from the Internet or an ISDN-PRI plugged into a PBX is a higher layer issue.

Thus, it was always an I.Q. test to see who would refer to "DSL service." I understand why the phone companies did it with their customers because end users can't be expected to know that level of detail. But any telco executive, engineer, marketing puke, VC or Biz Dev person who talks within the industry about "DSL service" is, by definition, an idiot.
dljvjbsl 12/5/2012 | 1:59:36 AM
re: Mobile Skype: Quality Issues? (sorry, encryption occurs above TCP),


In an IPSEC tunnel?
PO 12/5/2012 | 1:59:38 AM
re: Mobile Skype: Quality Issues? [283,dljvjbsl]:The provider will obviously have policies that determine the permissible calling rate for any particular subscriber. The zombie will bump up against that limit at the provider proxy and will not affect the overall network.

This reminds me very much of standard traffic shaping and policing. I do not see why the same effects are not expected in this case.


I've rarely seen a better description of how and why DDoS attacks succeed. Any individual source falls within acceptable and normal behavior. Collectively, their target is neutralized.

It's not that the security protocols don't do what they're designed to do: they do exactly what they're designed to do. The problem is that they can't do everything we want them to do.

Even if the BGP susceptibility (which is huge) were cryptographically protected (sorry, encryption occurs above TCP), it is important to remember that any analysis of protection is given with the caveat of what we know so far about the algorithms: a new algorithmic weakness could be discovered tomorrow.

Where does this lead? We need to look at security as a package, and never pretend that just putting a lock on the door solves the problem.

VoIP technology is fine, but the protocols are not ready to be used across a subscriber interface in an all-out deployment.
PO 12/5/2012 | 1:59:38 AM
re: Mobile Skype: Quality Issues? The end points are limited in their ability to handle this problem. This is a major reason why almost all of the nearly 500,000 ISDN-based video conferencing systems sit unused. These systems are notorious for QoS problems of various kinds, and those problems trace back to concantenation. The PSTN is great for individual 64 Kb/s streams, but don't dare rely on binding those channels together the treaing the result as a river.

Let's not blame the network for the failures of ISDN marketing, pricing, and adoption. The worst that can be said about ISDN is that engineers, not consumers, decided what the data rates should be. And that having to recondition lines that were longer than the network operators thought added to ISDN's cost burdens. But a lot was learned.

Similarly, a T1 circuit was always priced relative to the value of a DS0: the cost of channelized bandwidth was kept high because of its dual-use nature. Data services over DSL created a new pricing opportunity because the voice (PBX trunking) revenues were not (directly) threatened by DSL lines.
dljvjbsl 12/5/2012 | 1:59:39 AM
re: Mobile Skype: Quality Issues?
Experts previously said such attacks could take between four years and 142 years to succeed because they require guessing a rotating number from roughly 4 billion possible combinations


This was not a cryptographic issue according to the expert commentary. According to the experts, it was the guessing of the sequence number in the TCP stream.

The hack was to realize that with large TCP windows that this would be relatively easy. Thus the proposed measure of reducing the TCP window to only that which is needed. The experts also say that the BGP protocol is vunerable because of its long lived connections giveng more time for the hacker to make his guesses.
technonerd 12/5/2012 | 1:59:41 AM
re: Mobile Skype: Quality Issues? What technonerd does not seem to realize is that this bug is another manifestation of the in band control issue.
What dljvjbsl and the other data propellerheads do not seem to realize is that the dumb general public pays the bills. They don't want to know how it works, they just want it to work. A phone = a refrigerator = a car = an oven = a dishwasher. It is an appliance.

The public wants to know that if they pick up the phone at 2:30 in the morning they will hear a dial tone, not an excuse they can't understand. The computer-oriented shysters of California don't seem to get this. They continue to advertise things that don't work in terms that people can't understand. When this is pointed out, they become arrogant.

I have some news: The next wave is going to belong to a humbler bunch of people. I think it will be companies from India, but who knows maybe the Chinese will do it. But American companies, especially the ones from Silicon Valley, will miss the boat because they are way, way, way too full of themselves.

They remind me of the airlines, which hover on the brink of insolvency after so many years of having been at war with their customers. You heard it here first, guys. The U.S. high-tech consumer market, from top to bottom, is ripe for innovation from foreign competitors.
stephenpcooke 12/5/2012 | 1:59:41 AM
re: Mobile Skype: Quality Issues? Do you know that that type of fraud is so prevalent that is is not permitted to make credit card calls to many countries. There is just too much fraud.

Actually, I did know about this, though not the specifics of 2600Hz.
stephenpcooke 12/5/2012 | 1:59:41 AM
re: Mobile Skype: Quality Issues? As quoted in post #294:

Experts previously said such attacks could take between four years and 142 years to succeed because they require guessing a rotating number from roughly 4 billion possible combinations. Watson said he can guess the proper number with as few as four attempts, which can be accomplished within seconds.

So, cryptographic systems, previously thought to be impenetrable, can be cracked in 4 guesses within seconds. VoIP control signalling is encrypted and in-band. How long before a hacker figures it out? How long before he publishes the solution? How long will it take to fix?
dljvjbsl 12/5/2012 | 1:59:43 AM
re: Mobile Skype: Quality Issues?
The questions that will be asked are how quickly was the workaround arrived at? How complete a fix is it? Is this just another kluge or has it been thought through so we won't have another firestorm on our hands next week? Given the relative isolation of the PSTN there was time to make the changes, in the case of IP it is less likely that this will be the case


The issues that you bring up are important and I hope that the routing experts who read this board will comment on the issue in tbat regard.

However there is one issue that I have with your commentary. The PSTN was not isolated from this bug. The Blue Box fraud required a complete re-design of the trunking protocols on the PSTN. The hacker magazine 2600 took its name from the 2600Hz signaling on trunks that made the PSTN vunerable.

The Red Box fraud required that the protocol used on coin phones be changed and more importantly a redesign of the coin phone itself. It was not a trivial thing to fix. Hackers could trivially recreate the coin signals sent by the coin phone.

Given a minimum of hardware knowledge and access to the coinphone loop, it is just as easy to do today. I suppose that credit card fraud on coin phones is so easy that no one bothers.

Do you know that that type of fraud is so prevalent that is is not permitted to make credit card calls to many countries. There is just too much fraud.
stephenpcooke 12/5/2012 | 1:59:44 AM
re: Mobile Skype: Quality Issues? The questions that will be asked are how quickly was the workaround arrived at? How complete a fix is it? Is this just another kluge or has it been thought through so we won't have another firestorm on our hands next week? Given the relative isolation of the PSTN there was time to make the changes, in the case of IP it is less likely that this will be the case. In the meantime how much damage will have been done to corporate reputations?
dljvjbsl 12/5/2012 | 1:59:44 AM
re: Mobile Skype: Quality Issues?
Technonerd's description of the networks aside, his opinion is more likely to be that of the general public than yours. The article was posted from a publicly accessible and well-read website. Your rebuttal site wasn't even mentioned in any of these articles nor was their take on it now being trivial to bypass


I take what you mean. The public can be swayed by sensationalist reporting in the press. That is a problem that faces modern society and is, of course, not limited to VoIP. You did a very valuable thing in alerting this community to the issue.

Never the less, it took me less than 5 minutes to find the expert assessment of this issue. I would be interested in hearing the views of the routing experts that read this board to see what their assessment of this issue is.

However technonerd's take that this issue was somehow related to VoIP is entirly woithout foundation. The issue potentially affected Email, IM, web surfing, business processes and so on. It would have been nice to see a considered opinion.

What technonerd does not seem to realize is that this bug is another manifestation of the in band control issue. Contral and data can share the same connection. This issue was the cause of the Blue Box and REd Box frauds that affected the PSTN. The solution is to move to out of band signalling in which control signals are given a private connection iaccessible to hackers In the PSTN world this was done by a separate network which culminated in SS7 siganling.

In the IP world, the same idea of a separate network is used but these are provided by crytographic means. Control is given its own encryted channel. Thus the IPSEC work around recommended is in reality the same solution as done in the PSTN
stephenpcooke 12/5/2012 | 1:59:45 AM
re: Mobile Skype: Quality Issues? dljvjbsl,

Technonerd's description of the networks aside, his opinion is more likely to be that of the general public than yours. The article was posted from a publicly accessible and well-read website. Your rebuttal site wasn't even mentioned in any of these articles nor was their take on it now being trivial to bypass. As mentioned in your post this vulnerability has been around for quite some time. The public will say something like "what if it hadn't been a researcher with some moral fiber that found this hole...? What if it had been found by a hacker and published immediately via some hacker's network?"

The rate of exploitation of holes on the Internet far exceeds the rate at which these things can be fixed in any realistic manner. With an isolated PSTN this was never really an issue.
technonerd 12/5/2012 | 1:59:45 AM
re: Mobile Skype: Quality Issues? Depending on how well they were designed, and how much the standards change in the meantime, may mean forklift upgrades and serious security holes until that forklift is used

Yeah, and we all know how much the RBOCs love forklift upgrades. This is one more issue that leads me to think that their "VoIP" offering will actually by CVoDSL, which is merely another way of sending PCM from someone's home to the PSTN.
dljvjbsl 12/5/2012 | 1:59:45 AM
re: Mobile Skype: Quality Issues?
We constantly get warnings like this these days, and that's the problem. Why on earth would we dump a cheap, secure, paid-for, high-quality voice network in favor of an IP format that's such a box of worms?


Did you read the expert commentary nabout this issue at the URL that I posted or are you just spouting off? The issue is specific, well known and remedial.
technonerd 12/5/2012 | 1:59:46 AM
re: Mobile Skype: Quality Issues? Researchers uncovered a serious flaw in the underlying technology for nearly all Internet traffic, a discovery that led to an urgent and secretive international effort to prevent global disruption of Web surfing, e-mails and instant messages.

We constantly get warnings like this these days, and that's the problem. Why on earth would we dump a cheap, secure, paid-for, high-quality voice network in favor of an IP format that's such a box of worms?
dljvjbsl 12/5/2012 | 1:59:46 AM
re: Mobile Skype: Quality Issues? Please note that the URL for the TCP vunerability quoted in the article is incorrect. The correct URL is:

http://www.uniras.gov.uk/vuls/...


It does look as if this is not in any way a showstopping issue. Several effective methods are available to cope with it.
dljvjbsl 12/5/2012 | 1:59:47 AM
re: Mobile Skype: Quality Issues? In response to the TCP vunerability report, this message was sent on Dave Farber's Interesting People (IP get it!!) List

==========================================

Hi Dave

This email seems un-necessarily alarmist: The vulnerability in question has to do with reseting TCP connections, NOT computers.

The principles behind that attack have been known for years. The main contribution of the author is noting that BGP is particularly vulnerable to this attack based on the long-lived nature of the applications TCPconnections.

There is a good write-up available at http://www.uniras.gov.uk/vuls/.... tm that documents precisely what is being discussed right here. This writeup also notes some relatively easy work arrounds including

(A) Using IPSEC
(B) Reducing the size of the TCP Windows

Furthermore, the author notes that the TCP MD5 signature option is also an effective worrk arorund.





technonerd 12/5/2012 | 1:59:47 AM
re: Mobile Skype: Quality Issues? A correction: CIF tops out at 288 lines per frame and 360 pixels per line. QCIF tops out at 144 lines per frame and 180 pixels per line.

Neither one of these is very satisfying. In a word of Hi-Def and even ordinary digital TV via satellite, these rates will deliver a crappy looking picture. It just amazes me that the video conferencing world would hobble itself like this when there is almost literally a pot of gold waiting for the first guy who says, essentially: Screw this stuff, I'm going to give people a great video conference using technology that's been around for years.
stephenpcooke 12/5/2012 | 1:59:47 AM
re: Mobile Skype: Quality Issues?
Top News
Tuesday, April 20, 2004
Internet Technology Vulnerable to Hackers


WASHINGTON (AP) - Researchers uncovered a serious flaw in the
underlying technology for nearly all Internet traffic, a discovery
that led to an urgent and secretive international effort to prevent
global disruptions of Web surfing, e-mails and instant messages.
The British government announced the vulnerability in core
Internet technology on Tuesday. Left unaddressed, experts said, it
could allow hackers to knock computers offline and broadly disrupt
vital traffic-directing devices, called routers, that coordinate
the flow of data among distant groups of computers.
``Exploitation of this vulnerability could have affected the
glue that holds the Internet together,'' said Roger Cumming,
director for England's National Infrastructure Security
Coordination Centre.
The Homeland Security Department issued its own cyberalert hours
later that attacks ``could affect a large segment of the Internet
community.'' It said normal Internet operations probably would
resume after such attacks stopped.
The risk was similar to Internet users ``running naked through
the jungle, which didn't matter until somebody released some
tigers,'' said Paul Vixie of the Internet Systems Consortium Inc.
``It's a significant risk,'' Vixie said. ``The larger Internet
providers are jumping on this big time. It's really important this
just gets fixed before the bad guys start exploiting it for fun and
recognition.''
The flaw affecting the Internet's ``transmission control
protocol,'' or TCP, was discovered late last year by a computer
researcher in Milwaukee. Paul Watson said he identified a method to
reliably trick personal computers and routers into shutting down
electronic conversations by resetting the machines remotely.
Routers continually exchange important updates about the most
efficient traffic routes between large networks. Continued
successful attacks against routers can cause them to go into a
standby mode, known as ``dampening,'' that can persist for hours.
Experts previously said such attacks could take between four
years and 142 years to succeed because they require guessing a
rotating number from roughly 4 billion possible combinations.
Watson said he can guess the proper number with as few as four
attempts, which can be accomplished within seconds.
Already in recent weeks, some U.S. government agencies and
companies operating the most important digital pipelines have
fortified their own vulnerable systems because of early warnings
communicated by some security organizations. The White House has
expressed concerns especially about risks to crucial Internet
routers because attacks against them could profoundly disrupt
online traffic.
``Any flaw to a fundamental protocol would raise significant
concern and require significant attention by the folks who run the
major infrastructures of the Internet,'' said Amit Yoran, the
government's cybersecurity chief. The flaw has dominated
discussions since last week among experts in security circles.
The public announcement coincides with a presentation Watson
expects to make Thursday at an Internet security conference in
Vancouver, British Columbia, where Watson said he would disclose
full details of his research.
Watson predicted that hackers would understand how to begin
launching attacks ``within five minutes of walking out of that
meeting.''
On the Net:
U.K. National Infrastructure Security Coordination Centre:
www.niscc.gov.uk
Homeland Security cyberdivision: www.us-cert.gov





Help - Site Help Map - Advertise with Us - Add Site - What's New - What's Cool

Copyright -¨ 2000 Netscape. All rights reserved. Terms of Service | Privacy Policy
technonerd 12/5/2012 | 1:59:48 AM
re: Mobile Skype: Quality Issues? "... the PSTN cannot reliably switch video because it can't reliably concantenate DS-0 streams." Why does the switch have to concatenate DS0 streams. The end-points can use inverse muxes to handle this problem.
The end points are limited in their ability to handle this problem. This is a major reason why almost all of the nearly 500,000 ISDN-based video conferencing systems sit unused. These systems are notorious for QoS problems of various kinds, and those problems trace back to concantenation. The PSTN is great for individual 64 Kb/s streams, but don't dare rely on binding those channels together the treaing the result as a river.

This is a real weakness of the PSTN, and in all of my shit-missile launchings I have made no bones about it. The PSTN is terrible at video conferencing; it's a huge opportunity for IP, but to take advantage of it the IP crowd needs to yank their heads out of their assholes and re-examine the H.xxx standards designed for the limitations of the PSTN.

Until and unless they do this, video conferencing will remain stalled. The way to go about it is for some innovator -- as opposed to a subsidiary of a fraud-seeking, group-thinking California venture capitalist -- to put his superior system out onto the market, in partnership with a data network that will guarantee end-to-end QoS.

End-to-end QoS, at least on a limit basis and maybe more, really isn't that hard. But I think I'll leave some mystery on the topic. Surely the engineers of California know how to arrange it. If not, I suspect there are some in India and China who'd be happy to try.


"... they capped video at about 512 Kb/s because that's the top rate for concantenated ISDN." Is this really true? I didn't think 512 is one of the ISDN rates; on the other hand, ISDN allows for 1536/2048.
ISDN systems use multiple 64 Kb/s channels. The typical is 6 channels for video = 384 Kb/s but it can go higher. The H.261 and H.324 standards impose the CIF and QCIF formats for video conferences. No matter how much bandwidth you have, your resolution tops out 288 lines/frame and pixels per line if you're using an H.xxx device -- and all VC devices are H.xxx compatible.

These standards were designed to operate in the typical ISDN 6-channel concatenated mode. I wrote 512 Kb/s (eight 64 Kb/s channels) because I recalled hearing it from the videoconferencing guy who I trust the most. But I might have gotten that wrong. Maybe it was 384 Kb/s that I should have used, and frankly I'm a bit confused on that detail.

Yes, ISDN video conferencing systems can use more bandwidth, but CIF and QCIF will make it superflouous. The way to get past this is to dump CIF and QCIF in the IP realm and allow compatible video conferencing systems to take full advantage of available bandwidth. I think some military systems already do this, but I don't think widely available civilian systems do.
aswath 12/5/2012 | 1:59:49 AM
re: Mobile Skype: Quality Issues? technonerd:

In two different posts you have remarked about PSTN switching structure that I would like to comment.

In Msg. 288 you say, "... the PSTN cannot reliably switch video because it can't reliably concantenate DS-0 streams." Why does the switch have to concatenate DS0 streams. The end-points can use inverse muxes to handle this problem. By the way, analogous to DS0 switching, IP has MTU constraint; depacketization and jitter buffer handling are similar to inverse mux processing.

In Msg. 289, you say that, "... they capped video at about 512 Kb/s because that's the top rate for concantenated ISDN." Is this really true? I didn't think 512 is one of the ISDN rates; on the other hand, ISDN allows for 1536/2048.
stephenpcooke 12/5/2012 | 1:59:50 AM
re: Mobile Skype: Quality Issues? dljvjbsl wrote:

Secondly there are jsut too many variables for any one spec to get it all right.

Here we agree. Which is why it is confusing that some people will make audacious claims as to the notion of DoS (or any other type of) attacks not having a hope in hell (paraphrased) in a VoIP world. How can anyone be so sure? I will give those who are working on this credit for doing their best but unless God himself designs the network or the equipment that runs it, there will be problems.

In the end it doesn't matter how well the standard is written with regards to its approach to anything. I have mentioned many times on these boards that, as a VP in an independent test house, I saw NO, not one, fully standards compliant product. I will add that the standards that my company was testing to were quite mature and there were very few questions as to the exact meanings of the various requirements. Try that with standards that are nowhere close to being mature. First to market often wins the biggest portion of the market. To be first to market you can't wait until the standards are set in stone. This means that, even for new technologies there is a legacy problem to deal with. Depending on how well they were designed, and how much the standards change in the meantime, may mean forklift upgrades and serious security holes until that forklift is used.

I am sure the standards, such as they are, have benefitted from your knowledge dljvjbsl.
technonerd 12/5/2012 | 1:59:50 AM
re: Mobile Skype: Quality Issues? Check my words. I did not say that you were BobbyMax. I said 'Technonerd as BobbyMax.'

Aha! Hoist on my own petard (the close reading of words). Well, I thought Bobby Max was pretty good on a lot of things, except when he'd do his ranting about Indians. I've never been able to relate to that whole issue. But he was spot-on when he discussed the corruption of Silicon Valley and corporate America in general.
technonerd 12/5/2012 | 1:59:51 AM
re: Mobile Skype: Quality Issues? One that that I cna tell you for sure is that the SIP insiders take the issue of trust in these networks very seriously. What I am describing about a structured VoIP network is what I picked up from these sources.

One thing I would agree with is that real-time, full-duplex IP networks will need to emulate the deterministic nature of the PSTN if they hope to deliver reliability, QoS, security and scalability. This means that David Eisenberg's "Dawn of the Stupid Network" thesis was totally wrong.

As for "SIP insiders" taking the issue of trust "very seriously," that declaration and a buck will get you a cup o'joe at McD's. The reality is that IP was designed to be stochastic, and therefore security will always be a bolt-on.

I'm sure that, with a few decades of work, they'll jerry-rig something that will work most of the time. In the meantime, I think it would be a huge mistake to think that the PSTN is somehow "dead" or "dying" for moving PCM-formatted voice. The PSTN is the platinum standard, and this objective (hah! Technoboy!) reality drives the IP propellerheads absolutely crazy. After all, nothing hurts quite like the truth.


There are too many different types of networks and there are too many different vendor and customer needs, to do this. It is my opinion that this is where H.323 went wrong. It was too ambitious in thinking that it could define the needs of all networks for the forseeable time.
Another place where the H.xxx standards went grievously wrong is that they capped video at about 512 Kb/s because that's the top rate for concantenated ISDN. But getting one engineer, let alone a whole body of them, to admit a mistake is harder than convincing Bush that maybe the French and the Germans had a point.


Firstly the VoIP network will define new types of features beyond the traditional Call Forward type supplemental features. Secondly there are jsut too many variables for any one spec to get it all right.
I keep asking for specifics other than video conferencing. Technoboy gave us one valiant but inadequate example. No one else has been able to give any concrete examples of these new and fantastic applications. I am beginning to doubt they'll ever exist; not that California data people have ever just made it up as they go along to sell more shit that didn't work, doesn't work and won't work.

"Too many variables to get it right." Truer words never spoken! What always happens with the data propellerheads is they load their software and hardware with alleged "features" that haven't been tested, make no sense and do nothing other than eat up CPU cycles. It's a scam, and a whole lot of people are on to it.
technonerd 12/5/2012 | 1:59:53 AM
re: Mobile Skype: Quality Issues? This was an obvious bit of satire that illuminates the attitudes of one frequent poster who believes that the PCM [sic] network invented in the 1970s is the ultimate expression in technology and cannot ever be surpassed.

I will go waaaaaaaaaaaay out on a limb here and assume that I am the "frequent poster" to which this refers. First off, I use PSTN and PCM interchangably. Maybe I shouldn't, but when I do this it's because the PSTN carries bits formatted in Pulse Code Modulation. PCM is defined as follows, by Telecommunications Research Associates:

A technique for representing an analog waveform with a digital bit stream in which the analog signal is sampled at a fixed repetition rate and each sample is assigned a digital code word which represents the approximate amplitude of the sample. The sampling rate must be equal to or greater than twice the highest frequency present in the original analog signal. The amplitude range of the samples is broken into a set of discrete steps called quantization levels. The number of quantization levels depends on the number of bits assigned to each sample. For example, assigning 8 bits to each sample permits 256 quantization levels.

The quantization levels may be uniformly spaced or they may be spaced on a nonlinear basis. Nonlinear spacing is common in applications such as speech in which relatively small quantization steps are used for low-level signals and the spacing between quantization levels is gradually increased as the analog signal amplitude increases. This nonlinear approach allows the PCM coding to achieve a relatively constant signal to quantization noise ratio across a wide range of input signal amplitudes. The telephony standard for PCM in North America is called Mu-law PCM. The corresponding European standard is called A-law PCM.


As for the PSTN being "the ultimate expression in technology (that) cannot ever be surpassed," this is a very inaccurate reading of my postings. Here are some relevant beliefs of mine regarding the PSTN:

- It was purpose-built for voice. Attempts to adapt it for other uses have met with mixed results. Modems do an acceptable job for slow-speed data connections, but the PSTN cannot reliably switch video because it can't reliably concantenate DS-0 streams.

- The PSTN remains the cheapest, best-quality and most reliable means of delivering voice. It is so good at it, in fact, that advocates for alternatives have actually attacked the PSTN for being too good, too cheap and too reliable.

- If the "V" stands for "voice," VoIP is a joke. It is less reliable. It delivers lower quality. It uses the same backbone as the PSTN for transport and thus saves no money. The sole appeal of VoIP to any carriers or users is rooted in the regulatory scheme, which can be readily manipulated to divorces prices from costs.

- Other than packetized, real-time, full-duplex video, which would indeed be a major leap forward in terms of function, value and revenue-generation, advocates of packetized communications can demonstrate no specific, broadly useful applications unique to IP networks.
aswath 12/5/2012 | 1:59:55 AM
re: Mobile Skype: Quality Issues? This was an obvious bit of satire that illuminates the attitudes of one frequent poster who believes that the PCM [sic] network invented in the 1970s is the ultimate expression in technology and cannot ever be surpassed.

I apologise for being naive because the satire was lost on me. But I am still searching for a feature that VoIP can offer, but PSTN with ISDN type signaling can not. I will tell you why I am not hopeful.

VoIP does not care how the media is delivered; that means, it could come via B-channel as well. So the real distinguishing part is the signaling protocol. But then I can isomorphically map SIP to Q.931, save one. Because of IP transport, SIP can be addressed to any end-point, where as because of LAPD, Q.931 can be addressed only to a single switch. So how is that one can realize a new feature with SIP but can not with ISDN?
dljvjbsl 12/5/2012 | 1:59:55 AM
re: Mobile Skype: Quality Issues?
you seem to have a lot of inside information regarding SIP developments, perhaps you are contributing to them. Would this make you an objective commentator on the subject? It also seems that you are making assumptions about the providers of these services as justification for the presented scenarios not occurring. From this one can extrapolate that this is a valid weakness that is outside the jurisdiction or control of the SIP or VoIP standards


I hve attended the SIP and related working group meetings and contributed to their mailing lists.

One that that I cna tell you for sure is that the SIP insiders take the issue of trust in these networks very seriously. What I am describing about a structured VoIP network is what I picked up from these sources.

The SIP specs are focused on providing what is needed for what they do. The developers of the spec, quite rightly, have not attempted to define every aspect of the VoIP netowrk. There are too many different types of networks and there are too many different vendor and customer needs, to do this. It is my opinion that this is where H.323 went wrong. It was too ambitious in thinking that it could define the needs of all networks for the forseeable time.

As I have said before, this is one of the reasons that the SIP developers are not defining feature behaviours. Firstly the VoIP network will define new types of features beyond the traditional Call Forward type supplemental features. Secondly there are jsut too many variables for any one spec to get it all right.
stephenpcooke 12/5/2012 | 1:59:56 AM
re: Mobile Skype: Quality Issues? The provider will obviously have policies that determine the permissible calling rate for any particular subscriber. The zombie will bump up against that limit at the provider proxy and will not affect the overall network.

dljvjbsl you seem to have a lot of inside information regarding SIP developments, perhaps you are contributing to them. Would this make you an objective commentator on the subject? It also seems that you are making assumptions about the providers of these services as justification for the presented scenarios not occurring. From this one can extrapolate that this is a valid weakness that is outside the jurisdiction or control of the SIP or VoIP standards efforts. Each of the presented scenarios is however predicated on the existance of VoIP technology.

Tha approach of taking shots at the stability of the existing PSTN begs the question, given the presented scenarios, is VoIP going to cause the PSTN to be better than it already is (ie: actively plug any existing holes) or is it going to cause additional weaknesses?
dljvjbsl 12/5/2012 | 1:59:57 AM
re: Mobile Skype: Quality Issues? How about call forwarding, transfer, and three way calling? Surely these new features make VOIP a viable service to contend with the PSTN.


New? *grin*

This was an obvious bit of satire that illuminates the attitudes of one frequent poster who believes that the PCM [sic] network invented in the 1970s is the ultimate expression in technology and cannot ever be surpassed.
dljvjbsl 12/5/2012 | 1:59:57 AM
re: Mobile Skype: Quality Issues?
This sort of trust model suggests another DoS attack: a subscriber can violate their provider's abuse policy, get their provider locked out of the network, and disappear. As has been pointed out, because of zombies (or implementation bugs) the subscriber doesn't even need to be aware of their abusive behaviours.


The provider will obviously have policies that determine the permissible calling rate for any particular subscriber. The zombie will bump up against that limit at the provider proxy and will not affect the overall network.

This reminds me very much of standard traffic shaping and policing. I do not see why the same effects are not expected in this case.

dljvjbsl 12/5/2012 | 1:59:58 AM
re: Mobile Skype: Quality Issues?
My personal issue with SIP is that it's an immature technology. In five years when you can look up all the message sequence diagrams for feature interaction in RFCs, it'll be suitable for mass deployment. Until then, it's really not a 100% baked solution and every vendor has their own home-brew methods for doing features.


It is a deliberate policy at the IETF to state that the feature MSCs are for information only. There is absolutely no intention, or at least there was absolutely no intention and should be no intention, to specify these older supplemental features.

I recall the 50 and more page documents describing the H.323 way of handling Call Forward on Busy and other straightforward features. This, in my opinion, was one of the primary reasons that H.323 foundered. Instead of working on features that would justify a broadband system, they specified features that could be done with conventional systems.

SIP and SIPPING are working on the new types of features that will justify the investment in IP telephony.
PO 12/5/2012 | 2:00:03 AM
re: Mobile Skype: Quality Issues? [#275:dljvjbsl] I was under the impression that a VoIP network would learn from the Email experience. There would be cooperation as in the blackhole effort to prevent the Email experience from reoccurring. So providers would not accept VoIP traffic from anyone on the Internet but only from trusted partners. Some mom and pop ISP which wished to set up a VoIP service would have to find a trusted carrier to vouch for its traffic. Its session set up traffic will be routed through the major carriers network. Otherwise it will not be terminated.

Oddly enough, the Mom & Pop is probably a more trustworthy traffic source: they have a better chance of actually knowing their customers, and of noticing and reacting to aberrant behaviours.

This sort of trust model suggests another DoS attack: a subscriber can violate their provider's abuse policy, get their provider locked out of the network, and disappear. As has been pointed out, because of zombies (or implementation bugs) the subscriber doesn't even need to be aware of their abusive behaviours. (So an innocent subscriber's good name is tarnished as well.)

Either way, even if GuruNetwork vouched for Mom&Pop, the chain of trust is weak: Mom&Pop may vouch for Son and for Daughter, who each in turn vouch for Friend. But they all have customers who are, to varying degrees, creating unwanted traffic. Do you no longer trust Son? Daughter? When this scales, there will always be someone new willing to let the abuser in if the money is right.

I don't believe I had mentioned this particular DoS attack earlier: there is a home-router vendor who released a product with a pre-configured NTP server address (and a few other poor implementation choices). Innocent consumers launched a DoS attack on the university hosting the NTP server, and significant effort was required to mitigate the effects (which still cannot be completely nullified). Unintended consequences, indeed.
alchemy 12/5/2012 | 2:00:05 AM
re: Mobile Skype: Quality Issues? technoboy writes:
How about call forwarding, transfer, and three way calling? Surely these new features make VOIP a viable service to contend with the PSTN.

New? *grin*

My personal issue with SIP is that it's an immature technology. In five years when you can look up all the message sequence diagrams for feature interaction in RFCs, it'll be suitable for mass deployment. Until then, it's really not a 100% baked solution and every vendor has their own home-brew methods for doing features. The legacy protocols like SS#7 and ISDN went through the same pains.
aswath 12/5/2012 | 2:00:06 AM
re: Mobile Skype: Quality Issues? Some mom and pop ISP which wished to set up a VoIP service would have to find a trusted carrier to vouch for its traffic. Its session set up traffic will be routed through the major carriers network. Otherwise it will not be terminated.

I had not considered this scenario because one of the main tenets of Stupid Network is disintermediation.
technoboy 12/5/2012 | 2:00:07 AM
re: Mobile Skype: Quality Issues? That's what they said about residential ISDN 20 years ago, too. My main objection to VoIP is the the "Voice" part of it. If it winds up being full-duplex video, then my objections basically vanish. As for the other promised "amazing applications," I note that repeated requests for examples continue to be unanswered except for one valiant but weak try by the Technoboy.

How about call forwarding, transfer, and three way calling? Surely these new features make VOIP a viable service to contend with the PSTN.
dljvjbsl 12/5/2012 | 2:00:08 AM
re: Mobile Skype: Quality Issues? If you want to think I am BobbyMax, go right ahead. But in case you're in an uncharacteristic mood for a hard fact, I'm not.

Check my words. I did not say that you were BobbyMax. I said 'Technonerd as BobbyMax.
dljvjbsl 12/5/2012 | 2:00:08 AM
re: Mobile Skype: Quality Issues?
Didn't the email problem come up only after "inter-domain" exchange was setup? Why wouldn't a similar problem come in trusting SIP proxies as well?


Previously I had a home Internet account with one of the large telcos. I used it to communicate with a colleague at a large university. One day, I discovered that my Emilas were bing rejected at the university server because the large telco's network had open mail relays. The telco was on the blackhole list.

I was under the impression that a VoIP network would learn from the Email experience. There would be cooperation as in the blackhole effort to prevent the Email experience from reoccurring. So providers would not accept VoIP traffic from anyone on the Internet but only from trusted partners. Some mom and pop ISP which wished to set up a VoIP service would have to find a trusted carrier to vouch for its traffic. Its session set up traffic will be routed through the major carriers network. Otherwise it will not be terminated.

This would seem to be a tractable way of handling the NAT problem as well.

At least that is the impression that I have received from my discussion on the matter. Maybe I was just projecting my solution onto the discussions.
dljvjbsl 12/5/2012 | 2:00:08 AM
re: Mobile Skype: Quality Issues?
Oh really? Why do even the propagandists for VoIP talk about consumer willingness to trade the (allegedly) lower prices for VoIP for lower reliability and service quality? How come the VoIP geeks spend their time attacking five-nines?


VoIP networks are being sold to life and mission critical organizations. These customers such as police departments and hospitals seem to know what they want in reliability. They will certainly get it.

The SS7 network is a packet network controlling a transmission network. It seems reliable enough. Why a VoIP control packet network supervising a transmission network cannot be made just as reliable is an interesting question. Certainly as others have pointed out, the same solutions apply.

All in all this is just another BM example, isn't it.
technonerd 12/5/2012 | 2:00:09 AM
re: Mobile Skype: Quality Issues? Technonerd loves POTS/PSTN because the system is ever so reliable.

Not quite. I love the reliability of the PSTN, and can see absolutely no reason -- other than the regulatory end-run being plotted by the RBOCs and the FCC -- to replace it with VoIP.
technonerd 12/5/2012 | 2:00:09 AM
re: Mobile Skype: Quality Issues? But everything IP is where the future is. VOIP is just the begining!

That's what they said about residential ISDN 20 years ago, too. My main objection to VoIP is the the "Voice" part of it. If it winds up being full-duplex video, then my objections basically vanish. As for the other promised "amazing applications," I note that repeated requests for examples continue to be unanswered except for one valiant but weak try by the Technoboy.
aswath 12/5/2012 | 2:00:11 AM
re: Mobile Skype: Quality Issues? Msg #264: And also remember that SIP does not work like Email. There is a defined network with proxies which is completely unlike the Email network.

Didn't the email problem come up only after "inter-domain" exchange was setup? Why wouldn't a similar problem come in trusting SIP proxies as well?
lastmile 12/5/2012 | 2:00:11 AM
re: Mobile Skype: Quality Issues? BobbyMax was never an individual but Technonerd is.
I am a nobody in terms of the vast knowledge that technonerd has, yet I disagree with his assessment of where this industry is going. Technonerd loves POTS/PSTN because the system is ever so reliable. But everything IP is where the future is. VOIP is just the begining!
technonerd 12/5/2012 | 2:00:12 AM
re: Mobile Skype: Quality Issues? If you want to think I am BobbyMax, go right ahead. But in case you're in an uncharacteristic mood for a hard fact, I'm not.
technoboy 12/5/2012 | 2:00:12 AM
re: Mobile Skype: Quality Issues? Now Now Oxygen Thief me thinks you protest too much!!!!
technoboy 12/5/2012 | 2:00:12 AM
re: Mobile Skype: Quality Issues? PO: The early internet used private peering. Early email used private peering. UUCP was all about private peering. But private peering networks do not scale well.

And trusted third-party relationships only work when rules are agreed upon and respected.

If you have to compare this to the early internet implementations I would agree that scalability would be an issue but I do not believe that VOIP networks will require anything approaching the number of peering relationships that the public internet required. Also the nature of the relationships may be very different. I think the more challenging aspect will be to agree on similar SLA models for QOS across the various carrier domains.

Finally, at least some of the aspects of security are addressed outside of the SIP protocol itself. It will basically be a layered security model that will rely on both capabilities that are intrinsic to the protocol itself and external. Session controllers are one aspect of the solution that would be external but this function may be incorporated into softswitches or other elements in the future. Session controllers perform several tasks such as topology hiding which is an important control for things like DDOS attacks. These same devices can also provide capabilities for peering between carrier networks and enforcing some of the trust relationships that need to exist.
PO 12/5/2012 | 2:00:13 AM
re: Mobile Skype: Quality Issues? And also remember that SIP does not work like Email. There is a defined network with proxies which is completely unlike the Email network. I do not see the scaling issues are comparable

Email uses SMTP (POP3) servers as the analogue of SIP proxies. ISPs run their own servers, corporations run their own servers, and some smaller organizations and individuals run their own servers. Plus, there are competitive mail service providers as well.

In a competitive world, ISPs and telephony carriers would be expected to run their own SIP proxies, as well as corporations and many smaller organizations and competitive service providers.

The fact that email carries signaling and content along the same path improves the chances of solving the problems in email. But again, the scaling problems create faults in the scope of trust.

The current issue of Business Communications Review includes the quote, referring to VoIP, "The law of unintended consequences has yet to be repealed."
technoboy 12/5/2012 | 2:00:13 AM
re: Mobile Skype: Quality Issues? PO

OK in your previous post you stated that in a small network this might work which I interpreted to mean it would not work in a large deployment which I would consider most service provider networks to be.
dljvjbsl 12/5/2012 | 2:00:14 AM
re: Mobile Skype: Quality Issues? Today's outburst
dljvjbsl 12/5/2012 | 2:00:14 AM
re: Mobile Skype: Quality Issues?
Remember, all of the authentication, authorization, non-repudiation, etc could be applied to email servers and clients. But the relevant experts do not believe this would solve spam: the scaling issues break it


And also remember that SIP does not work like Email. There is a defined network with proxies which is completely unlike the Email network. I do not see the scaling issues are comparable
technonerd 12/5/2012 | 2:00:19 AM
re: Mobile Skype: Quality Issues? You can always choose not to read my posts if it so boring but then how would you spend your lonely nights. Perhaps by looking at girls in fishnet stockings
I'm a technonerd. I'd rather watch The Twilight Zone
PO 12/5/2012 | 2:00:19 AM
re: Mobile Skype: Quality Issues? I am curious as to why you think the closed model would not work.

Huh? I said it will work. And under other topics I have said that VoIP will be deployed, but behind their subscriber interfaces, not across them. I said the closed model will not scale.

The early internet used private peering. Early email used private peering. UUCP was all about private peering. But private peering networks do not scale well.

And trusted third-party relationships only work when rules are agreed upon and respected.

Remember, all of the authentication, authorization, non-repudiation, etc could be applied to email servers and clients. But the relevant experts do not believe this would solve spam: the scaling issues break it.

SIP has not addressed security sufficiently for use as a public protocol.
technoboy 12/5/2012 | 2:00:20 AM
re: Mobile Skype: Quality Issues? PO wrote:At the end of it all, we can look at a security model where carriers operate in a closed environment, prohibiting open competition but establishing a network of trust among themselves. And in a small enough world this could possibly be made to work.

I am curious as to why you think the closed model would not work. I reference the closed model in this context which is that the RBOC or ILEC provides the service on their network which is different from the model that say Vonage is using. Given that service providers are building these networks today with many looking to build down to the access network (which will be a challenge) why wouldnt private peering work in this environment. This has been the subject of a lot of discussion and it is an important point. Many who post on this thread assume that every provider will want to leverage the public internet (which in my view is an ambiguous term) while most service provider are steadily building their own networks and adapting legacy traffic and VOIP over that network. This makes sense because it allows for the carrier to carefully control traffic that enter and exits their networks.
technoboy 12/5/2012 | 2:00:22 AM
re: Mobile Skype: Quality Issues? Oxygen thief (formerly known as technonerd) wrote:If you want to have a fetish, might I suggest girls in fishnet stockings and spike heels? Your objectivity fetish is not only pointless, it's boring.

Technoboy responds: You can always choose not to read my posts if it so boring but then how would you spend your lonely nights. Perhaps by looking at girls in fishnet stockings
dljvjbsl 12/5/2012 | 2:00:23 AM
re: Mobile Skype: Quality Issues?
By the way, some appeared to be confused about an earlier comment: within IP, identification is provided by the "from" IP address which as I stated is rarely verified. This security discussion is entirely above IP.)


As you state the issue of security in SIP has been addressed.

The validity of IP addresses is another matter entirely. It defeats the purpose of SIP in setting up a connection to provide an spoofed IP address. The session could not beset up and no connection could be made on the voice or other data path.

It looks to me as if this provides a problem for a hacker. He'll have to make sure that the computer used in any attack cannot be traced back to him.

technonerd 12/5/2012 | 2:00:25 AM
re: Mobile Skype: Quality Issues? Why should others feel the need to pretend they are objective when they are clearly not
If you want to have a fetish, might I suggest girls in fishnet stockings and spike heels? Your objectivity fetish is not only pointless, it's boring.
PO 12/5/2012 | 2:00:27 AM
re: Mobile Skype: Quality Issues? The problem with public key authentication mechanisms is that you have to trust the root certificate. A single disgruntled employee can blow away your security by blabbing the keying material to a hacker board.


As a real question, why is this not a problem with banks? It would seem far more interesting for a hacker to rob a bank of millions than to make a bunch of annoying phone calls.

I would imagine that the solution to this is to make the keying material inaccessible to employees. They keys would be kept in a secure server and manipulated only rarely.


There are a few ideas getting mixed up in this part of the thread. (Hmm. None of us have mentioned the actual topic recently. Oh well.) First, there is the question of authentication and nonrepudiation. Second, there is the question of privacy and secure transmission.

I do not have a digital certificate issued by my bank. I have a card number and a PIN number. The use of public key encryption comes in primarily to provide privacy as I exchange my identity with my bank: it is difficult for anyone else to eavesdrop on my credentials.

To a lesser extent, I could validate the bank's certificate to confirm they are indeed my bank. But let's say their cert was signed by Thawte. My method of verifying that Thawte is who they say they are is the same as I would have used to verify that my bank is who it says it is.

SIP Best Current Practices (RFC3665) has been cited, and suggests HTTP Digest Authentication (RFC2617). But there is an acknowledged susceptibility to Man In The Middle attacks, and it is unclear what the results would be to a replay attack. So we've put a lock on the front door, but the thief can still knock the door down.

Adding SSL can provide privacy but also adds complexity.

At the end of it all, we can look at a security model where carriers operate in a closed environment, prohibiting open competition but establishing a network of trust among themselves. And in a small enough world this could possibly be made to work.

Or we can look at a security model where there is open competition, and carriers can exist in parts of the world where the technology we're talking about is illegal.

All of this also relies on well-behaved implementations, and a reasonably 'clean' and well-behaved network. It is not clear that either will exist.

(By the way, some appeared to be confused about an earlier comment: within IP, identification is provided by the "from" IP address which as I stated is rarely verified. This security discussion is entirely above IP.)
technoboy 12/5/2012 | 2:00:28 AM
re: Mobile Skype: Quality Issues? Oxygen Thief

Everyone has a right to their opinion and you certainly are not shy about expressing your view. Why should others feel the need to pretend they are objective when they are clearly not. At least you have the gnads to be up front and express your opinions.
technonerd 12/5/2012 | 2:00:29 AM
re: Mobile Skype: Quality Issues? All these references to late night TV shows makes me think you are very lonely.
Funny you should say that because I recently discovered a high-definition TV network that shows newly produced Twilight Zones with a black guy as the moderator. He does a good Rod Serling.
technonerd 12/5/2012 | 2:00:29 AM
re: Mobile Skype: Quality Issues? You are masquerading as an objective poster but you are clearly biased in your statements.
What's wrong with having opinions? 'boy, you're obviously slanted in favor of VoIP. Having a slant is o.k.; the issue is whether someone has any intellectual honesty. I find you to be intellectually honest even if you are misguided. Having an opinion is a good thing, not a bad thing.
technoboy 12/5/2012 | 2:00:30 AM
re: Mobile Skype: Quality Issues? Dear Oxygen Thief:

I think you watch too much TV!!!! All these references to late night TV shows makes me think you are very lonely. So sad!!!
technoboy 12/5/2012 | 2:00:30 AM
re: Mobile Skype: Quality Issues? Dear Oxygen Thief:

I think you watch too much TV!!!! All these references to late night TV shows makes me think you are very lonely. So sad!!!
technoboy 12/5/2012 | 2:00:31 AM
re: Mobile Skype: Quality Issues? To Oxygen thief and Mr Cooke:

I think you two should get married!!!!
technoboy 12/5/2012 | 2:00:31 AM
re: Mobile Skype: Quality Issues? Mr Cooke:

You dont have to agree with me and thanks for the troubleshooting 101 lesson. It is interesting that you can make a statement such as knowing more than 99% of the general populace without actually knowing them. That is the definition of arrogance. Again you didnt answer the original question. But that is OK because I dont expect you to.
Actually I dont have to convince you of anything and I would not attempt to do so. You are masquerading as an objective poster but you are clearly biased in your statements. At least technonerd (now known as Oxygen thief), has the gnads to make his position clear ( wrong but clear).
technonerd 12/5/2012 | 2:00:34 AM
re: Mobile Skype: Quality Issues? The success of companies like Vonage and the value they deliver is proof.
Oh, the bubble reblows itself. Vonage makes no money. It never will. The only "success" it will deliver is if the VCs can shove it out the the mutual funds and pension funds at some outrageous markup. You know, like the data CLECs, millimeter wave radio carriers and fiber carriers that went bankrupt.


So is the case with VoIP and i am sure you will be there 3 years down telling people how great VoIP is
Time will tell, won't it? Here's what I think will happen in 3 years: Basically nothing. But in some longer period of time, if the FCC keeps going down the current regulatory road we could have something that either is VoIP or claims to be VoIP.

Either way, it will be unregulated and controlled by a monopoly in most cases and an oligopoly in some cases. Which means that prices will be high and reliability will be low.

If we're lucky, the RBOCs will simply lie about their VoIP and it will be CVoDSL running through the Class 5s and into the PSTN. The prices will increase, but at least it will work.
technonerd 12/5/2012 | 2:00:35 AM
re: Mobile Skype: Quality Issues? You still don't get it technoboy, I am a consumer and a voter and you are going to have to convince me that VoIP is better than what I already have.
Sadly enough, Steve, this might wind up being naive. The FCC may well make VoIP so attractive that the phone companies will switch to it whether we like it or not. And once that happens, we'll have the worst of all worlds: An unregulated monopoly charging astronomical prices for telephones that don't work as well as the cheap, efficient, reliable ones they replaced.


Talk directly to the people doing any maintenance work, not the managers.
Truer words were never spoken! If the people who ran phone companies would get into a van and spend a couple days with the guys who fix stuff ...

The California IP data geeks make the telephone managers look diligent. The geeks like the Technoboy regularly have their heads in one of two places: Either up their asses or in the clouds. It's just phenemonal what bad solutions the data people provide. In a fairer world, we'd hold Stalinist show trials for them.
technonerd 12/5/2012 | 2:00:36 AM
re: Mobile Skype: Quality Issues? I have missed you these last few days. I have decided to give you a new name for the board. From now on I will call you Oxygen Thief. It is appropriate given the fact that you continue to spew forth all these ridiculous statements spun as fact and provide little meaningful input while taking up alot of air in the process.

I am thinking of an old Twilight Zone episode. The gambler dies and goes to heaven. It is a casino where he wins the jackpot every time he plays. It dawns on him that he has actually gone to hell.

'boy, would you please improve your retorts?

technonerd 12/5/2012 | 2:00:36 AM
re: Mobile Skype: Quality Issues? So go educate yourself than come back and have an OBJECTIVE debate

Here is the objective truth: VoIP's performance is, and will be for a long time (maybe forever), a fraction less than 1 of the PSTN's.
technonerd 12/5/2012 | 2:00:37 AM
re: Mobile Skype: Quality Issues? The new VoIP netowrk is being designed to the same reliability specs as the old digital network and the old electromechanical network before it.

Oh really? Why do even the propagandists for VoIP talk about consumer willingness to trade the (allegedly) lower prices for VoIP for lower reliability and service quality? How come the VoIP geeks spend their time attacking five-nines?
stephenpcooke 12/5/2012 | 2:00:38 AM
re: Mobile Skype: Quality Issues? Hi Technoboy,

Anyone who disagrees with you doesn't necessarily need to be 'corrected'. Disagreement is how we live, it keeps us sharp, and, if constructive in nature, can make us much better. In these posts I have tried to provide the benefit of my experience in telecom, as it currently exists, testing (there are few on these boards with my experience here), and network debugging. Sit in front of a customer whose network isn't working as they expect it to be, figure out what is wrong as quickly as possible (the technology or the expectations), and position the response without losing that customer. This turns the boys into men in a hurry.

I won't turn the fair is fair thing on you and answer the questions that I asked of you, I think we all have a good idea what the answers would be.

VoIP technology may be inevitable. That does not mean that it will not be the most costly technology introduction in history in people terms. You still don't get it technoboy, I am a consumer and a voter and you are going to have to convince me that VoIP is better than what I already have. How many consumers are going to accept the arguement that 'if you haven't been in a production VoIP lab in the last two years you have nothing to say on the topic?' I am more technically competent than 99% of the general populace and more of a risker as well. If you can't convince me how are you going to convince them?

Here are some tips for network debugging from my experience, you can take them or leave them as you wish:
- make your log entries as verbose as you can. This gives the people reading them the best chance possible of minimizing the impact of the problem.
- make your log dumps as verbose as you can, often the key lies in what is NOT there as much as what IS there.
- draw a map, as best as possible, so that you get the configuration straight before you start. Include everything that you know about each node so you can get a feel for the situation.
- LISTEN more than talk, especially to the techs. Talk directly to the people doing any maintenance work, not the managers. The techs may have done something that wasn't company policy and may get fired for it. Try to talk to them without the manager and, if you can save that person's job you will have taught them a valuable lesson and gained an internal advocate in the future.
- build in testability features from the outset. Traffic generators, packet verifiers, etc. can be worth the time to implement in no uncertain terms. Trade-offs will always happen in scheduling 'revenue-generating' features with troubleshooting features but they may save you in the end. Do this with an eye to using these features in the field and not just the lab.
- Who are the best designers? They are not the guys with the best marks from the best schools; I have known technicians that were more useful than PhD's. They are not the 'socialites' that drive fancy cars and are good with the opposite sex; their eyes are often on the wrong ball(s). In my experience the best designers are those who have done some time in the trenches of field support. Designers in their own cocoon try their best but often have no idea what their products will be used to do or the situations that they will be put in.
- Every situation is different, not just for you but for the people at the other end as well. Being able to think on your feet and be courteous at the same time can save the day in more ways than one.
- Often you have no idea what is causing the problem, sometimes it doesn't happen repetitively or it is very sporatic. Design your system so that you will have sufficient information to diagnose those situations as best you can. If you can show your customer that you are making your best effort they may keep your company around for a while longer.
- Learn something from every problem. Keep your field support people up-to-date on solutions to problems and how they were diagnosed. Internal communications is a lost art. Those who are good at it are far more likely to succeed than those who are not.
- Be as detail-oriented as possible. Don't let people make off-the-cuff remarks without understanding EXACTLY what they meant. The turn of a phrase can have disastrous consequences.
- Make sure that you understand how every piece of information was obtained. What diagnostic tool was used in what circumstances.
- keep a timeline if possible via log timestamps or whatever you have available. This will give you key insights into the sequence of events that brought about the difficulties.
- use every piece of information that you have available to you, including external test equipment or logs/alarms from competitor's equipment. Pride in your equipment to the exclusion of everything else will almost certainly cost you your customer if nothing else.
- If something doesn't jive between multiple sources of data, get to the root cause as fast as you can. This will allow you to give a weighting factor to various datapoints in the future.
- If the problem is a 'known issue' you had better make sure that the fix is already planned for in development and that you have a story for your customer. You have no idea how quickly these things can spread from customer to customer so if you want your company to survive you have to minimize these as quickly as possible.

There you have it. There is much more but you will have to learn that on your own. Now I can go become a truck driver.
dljvjbsl 12/5/2012 | 2:00:40 AM
re: Mobile Skype: Quality Issues?
The problem with public key authentication mechanisms is that you have to trust the root certificate. A single disgruntled employee can blow away your security by blabbing the keying material to a hacker board.


As a real question, why is this not a problem with banks? It would seem far more interesting for a hacker to rob a bank of millions than to make a bunch of annoying phone calls.

I would imagine that the solution to this is to make the keying material inaccessible to employees. They keys would be kept in a secure server and manipulated only rarely.
dljvjbsl 12/5/2012 | 2:00:41 AM
re: Mobile Skype: Quality Issues?
The problem with public key authentication mechanisms is that you have to trust the root certificate. A single disgruntled employee can blow away your security by blabbing the keying material to a hacker board.


As a real question, why is this not a problem with banks? It would seem far more interesting for a hacker to rob a bank of millions than to make a bunch of annoying phone calls.

I would imagine that the solution to this is to make the keying material inaccessible to employees. They keys would be kept in a secure server and manipulated only rarely.
alchemy 12/5/2012 | 2:00:41 AM
re: Mobile Skype: Quality Issues? dreamer101 writes:
Hi dljvjbsl,
I think the proper sentence that you should use is

"The identity of the originator in SIP CAN be protected by a digital ceritficate"

It is optional. I believe fair amount of VoIP service out there do not use that. The protection is more based on the MAC address of the VoIP adapter being pre-registerred in the SIP server before SIP service is being used...

Dreamer


Nope. The MAC address only has local significance on a LAN. Once you traverse through a router, you don't have a clue what the MAC was.

SIP uses a challenge authentication scheme in the REGISTER sequence. If you scan the first few pages of the SIP best practices document (RFC 3665), you'll see an example message sequence diagram. The problem with public key authentication mechanisms is that you have to trust the root certificate. A single disgruntled employee can blow away your security by blabbing the keying material to a hacker board. Since your SIP device can be anywhere on the internet (you can take your SIP phone with you and use it in your hotel room), there's a huge vulnerability to theft of service. Once you know the keying material, it's as easy to clone a SIP phone as it is to clone an AMPS analog cell phone. As long as a hacker is using an anonymous IP address, they'll be difficult to catch.
alchemy 12/5/2012 | 2:00:42 AM
re: Mobile Skype: Quality Issues? dljvjbsl writes:
However I would also like to reiterate that E911 systems do fail. Locally Bell is having problems with the E911 system and does not know the cause. I have also read in the NY Times of persistent problems in the E911 system in New York.

The PSTN is not some magical network that does not fail. It is an engineering wonder but it is still a product of human engineering. It can and does fail


An E-911 tandem is just a class 5 office with a special software load and a bunch of specialized answering positions. The problem is that it is owned by the state/local government so many of them are hopelessly obsolete. Typically, the class 5 office is OK but the answering/dispatch positions are antiquated and understaffed. I live in a state with a new E-911 network and I pay a couple o' bucks a month in my phone bill to foot the bill. NYC could upgrade their network but there are too many fingers in the cookie jar for that to happen anytime soon.

The important point is that the PSTN was engineered so that it's unlikely to fail. All the switching components are supposed to be five 9's and all interconnections between those elements are supposed to be redundant with fast failure detect and reroute. The last 5 miles isn't five 9's but it doesn't need to be.

You can do the exact same thing for VoIP in the MSO network. On an MSO network, the new generation CMTSs from Arris, Motorola, ADC, and Terayon are carrier class boxes. (Note that Cisco is missing from the list and is in something of a product gap at the moment.) The managed IP network can be configured redundantly. You can use tried & true OSPF if you can live with 1 minute service outages as the network detects failures and reconverges. You can use MPLS with fast reroute if you want to junk your crappy Cisco edge routers and get something closer to what telco people think of as carrier grade. Or you could use carrier grade routers from Juniper and use a technology like Ethernet 802.3 link aggregation or resilient packet ring and get true 50 mSec path protection. Soft switches are redundant. Media gateways and signaling gateways are at least highly available if not redundant.
dljvjbsl 12/5/2012 | 2:00:43 AM
re: Mobile Skype: Quality Issues?
You can engineer a VoIP network where all essential services are kept local so 911 can survive an IP backbone failure. That's what the MSO VoIP deployments look like at the moment


This is a very good ovservation about the SS7 failure.


However I would also like to reiterate that E911 systems do fail. Locally Bell is having problems with the E911 system and does not know the cause. I have also read in the NY Times of persistent problems in the E911 system in New York.

The PSTN is not some magical network that does not fail. It is an engineering wonder but it is still a product of human engineering. It can and does fail
technoboy 12/5/2012 | 2:00:43 AM
re: Mobile Skype: Quality Issues? Some people find the truth amusing!!!
alchemy 12/5/2012 | 2:00:44 AM
re: Mobile Skype: Quality Issues? somebody wrote:
And what happens to the poor schmuck in East Dogshit, New Mexico who picks up his VoIP phone in the middle of the night when his daughter is having a seizure, finds out he can't call the ambulance and his daughter dies? What then? "Road kill on the information superhighway?"

dljvjbsl replied:
The SS7 network in the eastern US was shut down by a single bug is a rarely used portion of the software. So yes these situations can and will happen and emergency calls will fail.

Yeah, but 911 calls went through even though the SS#7 network failed since it's a different signaling path to the E-911 tandem.

You can engineer a VoIP network where all essential services are kept local so 911 can survive an IP backbone failure. That's what the MSO VoIP deployments look like at the moment.
sgan201 12/5/2012 | 2:00:44 AM
re: Mobile Skype: Quality Issues? Hi technoboy,

"Translated you have no practical experience with this technology... etc...."

Ha! Ha! Ha! Ha!

Thank you for providing me with my amusement of the day.. :-) :-)

Dreamer

technoboy 12/5/2012 | 2:00:45 AM
re: Mobile Skype: Quality Issues? Technonerd:

I have missed you these last few days. I have decided to give you a new name for the board. From now on I will call you Oxygen Thief. It is appropriate given the fact that you continue to spew forth all these ridiculous statements spun as fact and provide little meaningful input while taking up alot of air in the process.

It is also clear that you have no practical experiece with VOIP techology, hence you have no credibility. Now go take your Prozac and get to your happy place!!!!

technoboy 12/5/2012 | 2:00:45 AM
re: Mobile Skype: Quality Issues? Greetings StephenCooke:

Thanks for your reply:

Corrections follow:
I am not arrogant as demonstrated by my previous comments that the technology is still maturing. I have just asked that people gain a little perspective which has been ignored in this message list. I have also asked that people clarify their comments about such statements which has been ignored. Hence, my responses may seem arrogant but largely they are meant to make you more aware of the basic mistakes in the assumptions you have been making.

You didnt really answer my question and since you decided not to I will assume your practicial experience in working with the technology is little to none. Feel free to correct the assumption.

You shouldnt feel the need to correct the statement regarding a follower of technonerd. It is clear that you two tend to agree on the topic at hand.

Since you didnt answer my questions I will not be answering your questions which I think is only fair. Remember, it is your assumption that the technology will not work and will have serious vulnerabilities despite the way you have addressed your position as seemingly benign.

What would you consider a proper level of deployment of VOIP techology before you would sign off on it as viable for the consumer? Your statemetns that we are taking shots at the PSTN is a little silly. Just pointing out that the PSTN is not bullet proof but the thought that somehow VOIP makes the PSTN more vulnerable is another misconception. Dont expect people to educate you on how SIP Proxies and back to back user agents and session controllers and other components of VOIP technology work because on this board it is an invitation for the proverbial shit missile to be launched. So go educate yourself than come back and have an OBJECTIVE debate.
dljvjbsl 12/5/2012 | 2:00:46 AM
re: Mobile Skype: Quality Issues?
And what happens to the poor schmuck in East Dogshit, New Mexico who picks up his VoIP phone in the middle of the night when his daughter is having a seizure, finds out he can't call the ambulance and his daughter dies? What then? "Road kill on the information superhighway?"


The SS7 network in the eastern US was shut down by a single bug is a rarely used portion of the software. So yes these situations can and will happen and emergency calls will fail.

Software loads are large and complex. Adding new services can result in unexpected behavior. A lot of research on software verification and testing is done by the vendors and the carriers. The loads are probably as verified as anything can be but mistakes do creep through.

These systems do experience errored calls due to software. These are very low in number and there are clear specs on the allowable number but they do happen.

Hardware does fail. I remember helping my mother call my aunt. There were a set of local trunks that were obviously going high and dry with the network thinking that there was no problem. I called the Bell repair center and had great difficulty in finding anyone who knew about trunks or how to take a trouble report. I received instruction in waiting for dialtone. Eventually they did realize tha they had a major problem and were able to repair these trunks quite quickly.

Contrary to a great deal of what is said here, the PSTN is a large complex network filled with technology that is not perfectly designed and not perfectly maintained. It is a technological marvel but it does exhibit failures. It is not uncommon for E911 systems to fail with the obvious results.

The new VoIP netowrk is being designed to the same reliability specs as the old digital network and the old electromechanical network before it. The designers of each generation of technology have been able to meet the relaibility that people expect from the telpehone network. The VoIP networks that the carriers are deploying today will meet them as well.

dljvjbsl 12/5/2012 | 2:00:46 AM
re: Mobile Skype: Quality Issues?
Excuse me, but am I to assume that you have never created a CA and signed your own certificate? (Never mind that this is about the Internet's most popular security models, and not about how IP works.) Just because I might run a very small VoIP service doesn't prove whether I'm legit or rogue. Or that my "customers" are legit or rogue. And even if I get someone else to sign my certificate, so what? What CAs do you trust?


This is in reply to my comment on the statement:

Let's not forget that in the POTS world the CLID is added within the scope of trust. In the IP world, originator identifiers are added outside the scope of trust and are rarely verified. Forged source identifiers are already common.


Digital certificates are used to verify that a call originator is who he says he is. These ceterficates can be used to convey this trust through all of the carriers through which the call transits.

I suppose the major issue is that they are non-repudiatable. Only the originator of record of the call could have created them. This is entirely different from Email.

Which all being translated means that it is highly unlikely that any carrier is going to let your grandmothers VoIP phone initiate a global DDOS attack by calling every other phone (PSTN and VoIP) in the world.

As to which CA authorities will be trusted, this is up to the carriers. I would imahgine that the large carriers would set up their own authorities and also provide this service to the smaller mom and pop providers. It would seem to be a requirement for interconnection that would be rigidly enforced.
technoboy 12/5/2012 | 2:00:47 AM
re: Mobile Skype: Quality Issues? Hi Dreamer101

Translated you have no practical experience with the technology in the last two years as I suspected. Not that it will prevent you from spewing your thoughts on the subject. I have a suggestion. Wake up from the dream!!!!
technovc 12/5/2012 | 2:00:48 AM
re: Mobile Skype: Quality Issues? :-)))))
You seem to be really angry with new technology !!!

I agree VoIP has its pitfalls now BUT if you look at what carriers are eveluating (China tel, singtel, MCI etc) it is clear that they are going VoIP.

The success of companies like Vonage and the value they deliver is proof.

of course a guy in mexico may not get the best reliability on VoIP but so is the case there with PSTN or for that matter in Afganistan!!

I am sure there were sceptics like you when Graham bell invented the telephone , BUT the invention went on to become what it is today. So is the case with VoIP and i am sure you will be there 3 years down telling people how great VoIP is :-))))
Technovc

technonerd 12/5/2012 | 2:00:48 AM
re: Mobile Skype: Quality Issues? The fact is the internet is becoming all pervasive and VoIP is getting recognition as the Voice transport mechanism for this century
That's sloganeering, not analysis. It's typical of what passes for intellect among group-thinking Californians.


Give it a bit of time and it will become as sturdy as the PSTN.
Speaking of Californians and their inability to comprehend anything more complex than a sound bite, how about this: Serve no wine before its time.


Whatever we would say VoIP is here to stay
Like, uh, Fiber Channel, millimeter wave radio and Free Space Optics -- for starters?


people are going to deploy, make mistakes, learn , make mistakes again and learn again but in the process those that deploy and learn early will make/save a lot of money
And what happens to the poor schmuck in East Dogshit, New Mexico who picks up his VoIP phone in the middle of the night when his daughter is having a seizure, finds out he can't call the ambulance and his daughter dies? What then? "Road kill on the information superhighway?"


Look at all the telecom vendors , the alcatel`s, lucents, Ericssons etc they clearly see VoIP coming and hence the rush into this space.
I'm not sure about those things, but one thing I can see is an incoherent expression from someone who would have flunked my 5th grade English class. And when I see it, I shudder to think that such people might be in charge of the future.


Look at the tradeshows what do you see? VoIP everywhere.
Actually, what you see everywhere at trade shows is booth babes everywhere. I will forget more of those events than you will remember. I have seen more stupid ideas that never worked than I can possibly count.
PO 12/5/2012 | 2:00:51 AM
re: Mobile Skype: Quality Issues? Let us not forget that the mistakes of Email have not been propagated to SIP and VoIP. The identity of teh ortignator in SIP is protected by a digital certificate. It is impossible to forge in any practical manner. I would ahve to agree with Technoboy in this. There are some basic assumptions here about how IP works, that are wrong.

Excuse me, but am I to assume that you have never created a CA and signed your own certificate? (Never mind that this is about the Internet's most popular security models, and not about how IP works.) Just because I might run a very small VoIP service doesn't prove whether I'm legit or rogue. Or that my "customers" are legit or rogue. And even if I get someone else to sign my certificate, so what? What CAs do you trust?

Traditional telecom carriers have significant interconnection agreements. (And there are legal implications for fraudulent CLIDs.) Clearly you have a specific trust model in mind for SIP providers (and users) across the globe, but there are multiple models in play.
technovc 12/5/2012 | 2:00:51 AM
re: Mobile Skype: Quality Issues? Stephen,

I agree with you that VoIP is still not at the stage PSTN is and it will not be , it took a century for PSTN to get to where it is today . The internet itself is quite young compared to the PSTN and VoIP is probably only a few years old . The fact is the internet is becoming all pervasive and VoIP is getting recognition as the Voice transport mechanism for this century . Give it a bit of time and it will become as sturdy as the PSTN.

Whatever we would say VoIP is here to stay , lets not forget that , people are going to deploy, make mistakes, learn , make mistakes again and learn again but in the process those that deploy and learn early will make/save a lot of money.

Look at all the telecom vendors , the alcatel`s, lucents, Ericssons etc they clearly see VoIP coming and hence the rush into this space. Look at the tradeshows what do you see? VoIP everywhere. If there is so much interest and so many people working on it , this is only going to get better and better.



Technovc
technonerd 12/5/2012 | 2:00:51 AM
re: Mobile Skype: Quality Issues? Stephencooke you appear to be a follower of technonerds thought process so tell me have you been actively engaged in the deployment of VOIP technology either in the carrier or enterprise space in the last two years. Im just trying to get a sense of your background and experience with the technology. I noticed that some of these posts are still making some very basic assumptions about IP Communications that happen to be inaccurate.
Remember that old Star Trek episode starring the extraterrestrial drawf who turned out to be a friendly little fellow? Technoboy, to quite said drawf: Drink the tranya!
technonerd 12/5/2012 | 2:00:51 AM
re: Mobile Skype: Quality Issues? To be fair, I would often answer yes not because the product was in the shape I wanted it to be in, but because I knew the support groups were as prepared as possible to react to problems as they arose.
Hmm, so you never said yes because you knew it was what they wanted to hear and if you said no one too many times you might be fired? Nah, that never happens in Silicon Valley!
PO 12/5/2012 | 2:00:52 AM
re: Mobile Skype: Quality Issues? Ask your quality people if they feel that your product is 100% ready for deployment. Here is a hint, if they say yes, fire them.

To be fair, I would often answer yes not because the product was in the shape I wanted it to be in, but because I knew the support groups were as prepared as possible to react to problems as they arose. Of course, I was in design, not QA.
stephenpcooke 12/5/2012 | 2:00:53 AM
re: Mobile Skype: Quality Issues? Technoboy,

Interesting, I thought that I was the one who initiated several of these threads, I guess it must have been a subliminal thing. I think that I have expounded on my level of experience in this industry enough in the last few months on these boards. If you are interested you can look them up. You will also know that I have been talking about the side of the arguement that has to do with the exposure of the PSTN to the Internet, full of daily virus-updates and worms, which it has never been faced with before.

I have a lot of faith in those still left in the industry. Many of the best have been carved out and are now doing other things, as I will be soon. I am sure that you are working hard to make VoIP as good as you hope it can be. I submit that you are not even close to being unique in this. Every engineer who has ever designed a circuit or written a line of code has likely tried to do their best to make sure that they have covered all the bases. Even the best have made mistakes. If you feel that you and the rest of the VoIP world are better than those who have gone before, your arrogance will be your undoing. I have simply tried to illustrate some very small examples of what could, and will most likely be tried by individuals or even rogue countries, that haven't been fielded very well by zealots such as yourself.

Here are a few questions for you:
- How many outstanding trouble tickets (or whatever your company calls problems found in testing at one level or another) are there against your product at this very moment?
- Ask your quality people if they feel that your product is 100% ready for deployment. Here is a hint, if they say yes, fire them.
- If you feel that your technology is bulletproof, you should be willing to insure it against any kind of attack. Once you have offered this to any RBOCs come back and tell us how many of them took you up on it.

You have said in earlier posts that VoIP implementation is in the early stages (I'm paraphrasing). I submit that this implementation will take much longer to get traction and mass acceptance by consumers than most other technologies that already exist in the PSTN, because those technologies were introduced in a much more isolated and benign environment. Some of your compatriots have been taking shots at the current PSTN (isolated as it is), many of which are valid. Take a second and imagine how much worse it might get with malicious entities resident on the Internet. Ask any Internet security consultant how safe the Internet is. Ask them how they would hack into the PSTN via VoIP.

Your dedication is admirable, your faith is misplaced.
dljvjbsl 12/5/2012 | 2:00:53 AM
re: Mobile Skype: Quality Issues?
I believe fair amount of VoIP service out there do not use that. The protection is more based on the MAC address of the VoIP adapter being pre-registerred in the SIP server before SIP service is being used...


You may be right but why would anyone care about a MAC address. The proxy, registrar and UA can be anywhere in the network
dljvjbsl 12/5/2012 | 2:00:53 AM
re: Mobile Skype: Quality Issues? Spoofing the originators IP address will not do much f\good in VoIP. One needs to negotiate to set the call up. Thus a spoofed IP address will defeat the hacker's purpose. Email is entirely different and spoofed addresses ahve no real effect on the operation of sending a message
sgan201 12/5/2012 | 2:00:54 AM
re: Mobile Skype: Quality Issues? Hi Technoboy,
If the emperor has no clothes, it does not take a tailor to see that he is naked...

Dreamer
sgan201 12/5/2012 | 2:00:54 AM
re: Mobile Skype: Quality Issues? Hi dljvjbsl,
I think the proper sentence that you should use is

"The identity of the originator in SIP CAN be protected by a digital ceritficate"

It is optional. I believe fair amount of VoIP service out there do not use that. The protection is more based on the MAC address of the VoIP adapter being pre-registerred in the SIP server before SIP service is being used...

Dreamer
itisi 12/5/2012 | 2:00:54 AM
re: Mobile Skype: Quality Issues? The phone company that terminates a call may not be the one that originates it. Trust is necessarily transitive.

IP source addresses are easy enough to (reasonably) validate at the edge via Unicast RPF lookups - and damn well ought to be validated, though as you say are typically not. With enough legal/social/business pressure, one might expect that some day they will be. Rogue ISPs fall in the same category as rogue Telcos: possible, could do a lot of damage, but the sanctions, if sufficiently daunting, will likely inhibit them.
-------------
Let's not forget that in the POTS world the CLID is added within the scope of trust. In the IP world, originator identifiers are added outside the scope of trust and are rarely verified. Forged source identifiers are already common.
aswath 12/5/2012 | 2:00:55 AM
re: Mobile Skype: Quality Issues? Let's not forget that in the POTS world the CLID is added within the scope of trust. In the IP world, originator identifiers are added outside the scope of trust and are rarely verified.

I would like to bring to your attention a fact that may not be directly relevant for the point made in the post. Even in POTS, CLID may not be so trustworthy. If I am not mistaken, PBXs deliver CLID on the trunk interfaces and they can populate with a number that has no relationship to the trunk interface.
technoboy 12/5/2012 | 2:00:55 AM
re: Mobile Skype: Quality Issues? OK,

Its nice to know you can leave the board for a few days and pick up the conversation where it left off. Stephencooke you appear to be a follower of technonerds thought process so tell me have you been actively engaged in the deployment of VOIP technology either in the carrier or enterprise space in the last two years. Im just trying to get a sense of your background and experience with the technology. I noticed that some of these posts are still making some very basic assumptions about IP Communications that happen to be inaccurate.
dljvjbsl 12/5/2012 | 2:00:55 AM
re: Mobile Skype: Quality Issues?
Let's not forget that in the POTS world the CLID is added within the scope of trust. In the IP world, originator identifiers are added outside the scope of trust and are rarely verified. Forged source identifiers are already common.


Let us not forget that the mistakes of Email have not been propagated to SIP and VoIP. The identity of teh ortignator in SIP is protected by a digital certificate. It is impossible to forge in any practical manner. I would ahve to agree with Technoboy in this. There are some basic assumptions here about how IP works, that are wrong.
PO 12/5/2012 | 2:00:56 AM
re: Mobile Skype: Quality Issues? Be it noted that there's an asymmetry in DOS attacks in the different technologies. On the one hand, it's much easier to launch DOS in the IP space. On the other hand it's much harder to resist in the POTS space.

Let's not forget that in the POTS world the CLID is added within the scope of trust. In the IP world, originator identifiers are added outside the scope of trust and are rarely verified. Forged source identifiers are already common.

A lot of bandwidth has been given to discussing attacks on (carrier) servers. There are also opportunities for attacks on clients from rogue "network operators".
aswath 12/5/2012 | 2:00:57 AM
re: Mobile Skype: Quality Issues? Look at Pulver's company. They charge people for VoIP. There's a revenue stream.

As far as I know, pulver's FWD has no charge; as the name suggets, the directory service (that is all they offer currently) is free.
dljvjbsl 12/5/2012 | 2:00:58 AM
re: Mobile Skype: Quality Issues?
implying that a new VoIP company was looking to limit their revenue....? Companies do not care what credit card meets their authorization, just that one does. Once it has been authorized they are good to go.


It would also not take very much of this to put any real-world credit card over its limit. Additionally the fraud detection software at the credit card company would pick up these transactions
technonerd 12/5/2012 | 2:00:59 AM
re: Mobile Skype: Quality Issues? I thought that I was clear that it was very difficult to perform DoS attacks using modems (in a non-VoIP world) because the modems would actually have to dial someone's phone number
I must have misinterpreted your posting. Sorry about that.


Your discussion on channelized DSL and its marketing is fine except for the touted opex savings of VoIP networks. Your technology solution still uses POTS. Having said this, I like the approach much better than pure IP
I'm not recommending that they do this, but rather observing what has been deployed and predicting that this is how they'll go. As for opex savings of VoIP, all I'll do is focus on that word, touted. Incidentally, if the BOCs to the CVoDSL thing, be prepared for them to find a way to call it packetized voice. The real goal is to wriggle out of regulation.

Along those lines, remember: All telephone companies are evil, and any vendor or carrier will tell any lie they think they can get away with. These days, they've been getting away with enough to make me think they'll try to present CVoDSL as a revolution. Think of it this way: Who'll stop 'em?


Define flat-rate. If you mean $29.95/month, unlimited useage the problem just got much easier for hackers and much harder to track.
That's what I mean. All-you-can eat. Perhaps it just easier for hackers -- I don't know -- but you wouldn't have them fraudulently charging individual calls to a credit card, that's all.


all of which didn't have access to the PSTN until VoIP whether they were infected or not.
That's a great point. Thanks for bringing it up. I need to remember not to let my guard down. Thanks, Steve!
stephenpcooke 12/5/2012 | 2:00:59 AM
re: Mobile Skype: Quality Issues? technonerd,

I had to read your post, and mine a couple of times, I thought that I was clear that it was very difficult to perform DoS attacks using modems (in a non-VoIP world) because the modems would actually have to dial someone's phone number (ie: not AOL or some other ISP because the ISP's do not have the capability to ring someone's phone without VoIP technology). With the advent of VoIP, AOL could be used to launch such an attack from dial-up PCs.

Your discussion on channelized DSL and its marketing is fine except for the touted opex savings of VoIP networks. Your technology solution still uses POTS. Having said this, I like the approach much better than pure IP.

To be fair, I think most of the VoIP I've heard about is flat-rate stuff.

Define flat-rate. If you mean $29.95/month, unlimited useage the problem just got much easier for hackers and much harder to track. If you mean $.02/minute anywhere, anytime it is as I described.

I think DOS attacks are a function of insecure Windows-based computers, always on, attached to broadband connections, not the applications they're running.

all of which didn't have access to the PSTN until VoIP whether they were infected or not.
technonerd 12/5/2012 | 2:01:00 AM
re: Mobile Skype: Quality Issues? implying that a new VoIP company was looking to limit their revenue....? Companies do not care what credit card meets their authorization, just that one does. Once it has been authorized they are good to go.
Actually, I suspect that a new VoIP company is looking to be paid. If they get a swarm of signups, they'll have to wonder whether they'll ever be paid for these, i.e., whether the credit card companies won't be processing a whole lot lof chargebacks.
technonerd 12/5/2012 | 2:01:00 AM
re: Mobile Skype: Quality Issues? Most such taxation is based on gross revenues. Flat rate usage plans have no discrete revenues for the calls. Thus nothing to tax. All legal.
One way or the other, people will wind up paying extra money for real-time, full-duplex telecom. Whether each call is billed or whether they pay a flat rate is an irrelevant detail from the point of view of a taxing authority.

Look at Pulver's company. They charge people for VoIP. There's a revenue stream. If the government decides the tax this, there will be no real difficulty in doing so. As for Skype, et. al., i.e., peer-to-peer, if it ever scales up then people are going to pay extra to use it.

At that point, it will be taxed. If for any reason there's logistical difficulty in doing this, then the feds will simply impose USF, etc etc etc, on the broadband connection. People really ought to stop fooling themselves about this stuff somehow being free.
itisi 12/5/2012 | 2:01:00 AM
re: Mobile Skype: Quality Issues? Ah, but that rude shock will be legal/social rather than technological. The technological vulnerability of a circuit-oriented physical=logical endpoint is such that, without legal support, given endpoints could easily be rendered useless.

Pari passu, at some point DOS in the IP network may be the subject of similar sanctions. If ISPs and their end users were penalized for staging DOS attacks or held liable for the actions of zombies, defenses in the IP network would get bulletproof rather quickly.

Consider your own corporate IP net, if it's a large one. All the tools of DDOS can be brought to bear inside it. They won't be, because the perpetrators would be fired and the sysadmins would be obliged to clean up their network.

--------------------
You don't need a fancy DDOS zombie net to cripple most end points on the POS net.
Actually, you do. These days, kids doing an early 60s DoS attack on POTS will get a pretty rude shock, pretty fast.

technonerd 12/5/2012 | 2:01:01 AM
re: Mobile Skype: Quality Issues? You don't need a fancy DDOS zombie net to cripple most end points on the POS net.
Actually, you do. These days, kids doing an early 60s DoS attack on POTS will get a pretty rude shock, pretty fast.
technonerd 12/5/2012 | 2:01:01 AM
re: Mobile Skype: Quality Issues? DoS attacks via worms/viruses are meant to be generally undetectable by the computer user, at least until the event happens. Things like the phone being dialed when someone else wants to make a call, the phone going off-hook and modem sounds in the middle of the night, etc. would all require some form of investigation and potentially anti-virus-type response. The amount of logistical coordination required is unimaginable
DOS attacks are generally directed at a server. I'm not a data geek, but I would suspect that future DOS attacks against VoIP would be directed at a gateway or some other box essentially to the provision of the connection.

If I'm not mistaken, the DOS attacks work by hijacking other computers and using them to send traffic toward a particular point. I tend to doubt that the logistical side is all that unimaginable these days, given that all businesses and 20% of residences are connected to always-on broadband. It's the broadband users, not the dial-up ones, who are most vulnerable to having their machines hijacked.

There are "solutions" available to counteract this, i.e., firewalls, anti-spyware programs, virus checkers. But we really don't know how well they work. Anyone who spends any time investigating these issues will quickly find credible people on the Internet warning that all of the commonly used programs are wide open for attack.

This leads me to two observations:

1. Security in the IP world is a constant tug-of-war between 20-something hackers and providers of security solutions. This is because IP was, in some sense, designed from the ground up to NOT be secure, i.e., as a stochastic and connectionless means of moving content.

2. As a result of the underlying insecurity of public IP networking, end users must be more involved with the technology than users of PCM-based networking. It was designed with security in mind -- primarily for the purpose of billing for use of the network.

The only way IP phones will ever have a chance of being secure is if end-users grant the network operator and/or software provider a level of access to their computers that would also permit other intrusions, i.e., automatic downloading of anti-virus software could be accompanied by God-knows-what automatic uploading.

Oh yeah, I'm a paranoid freak. Not that anyone should ever have any reason to ever doubt Microsoft's intentions, diligence or skill. To allow the phone company to essentially control customer-owned CPE purpose-built for voice only is one thing, but to have an outside entity looking in on your computer all the time? I don't think so.

Even if users (other than myself) willingly granted this level of access to their machines -- which I don't think they will do -- the hackers might outrun whatever protection is provided, most likely in a future version of Windows. The bottom line is that, in the event it really gets popular, authentic VoIP is a rat's nest from a security standpoint.

As a result, I really don't think it will scale as VoIP. Too many issues. On the other hand, if the RBOC puts voice traffic onto DSL along with IP and then splits the voice onto the PSTN -- using Channelized Voice over DSL, which has been built into the new ADSL chips -- there will be none of these problems because the voice will never connect to the public Internet. In fact, it won't even be coded as IP, but who needs to tell anyone?

Why do this? To make a VoIP claim. They'd market the service as a new type of voice and claim it's just an application, which means it would be unregulated. The FCC is itching to kill off the Telecom Act on behalf of campaign contributors, so you just know they'd bless whatever the RBOCs tell them to bless. The Light Readings of the world know where their bread is buttered, and the general media are too stupid and lazy to figure it out.

There would be no VoIP at all under this architecture -- only a claim that would be supported by the p.r. departments at every RBOC, every Silicon Valley vendor, all the trade pubs, the general media and the FCC. In fact, the DSL connection would function very much like the pairgain, or DAML, devices of the '70s and '80s, with the one real difference being that it would also carry IP from the subscriber's computer to a router in the CO.

Only the IP would go to the Internet. The voice would have been coded as PCM and sent through the lower frequencies in a tunnel. Once the whole stream hits the DSLAM, the PCM bits are sent off to the Class 5. Call it a secret between friends.


The power of DoS attacks comes from the ability to infact tens of thousands or even millions of computers that can then act at the hackers command or at least their coordination.
Someone can correct me on this, but I doubt you can effectively build a DOS attack on computers connected via dial-up. Yes, there have been some spyware dialers installed on people's computers, but it's a pretty inefficient way to commit crimes. Not enough real money in it for the VCs and executives of Northern California, or the cyber-punks or whatever the hell they call themselves now.


VoIP calls can be charged to stolen credit cards without the owner being any the wiser, until they get the bill.
To be fair, I think most of the VoIP I've heard about is flat-rate stuff. And I can't quite see how credit card payments to a VoIP vendor would be any less secure than payments to anyone else.


Bottom line: VoIP has enabled DoS attacks where there was no realistic possibility before. It has opened Pandora's box to possibilities, good AND bad, that have never been considered before in the PSTN.
As intensely skeptical as I am about mass-market VoIP, and as gleefully willing as I am to launch shit-missile fusillades here, I'm not sure that I'd blame the VoIP application per se for DOS attacks. I think DOS attacks are a function of insecure Windows-based computers, always on, attached to broadband connections, not the applications they're running.

On the other hand, I'll be the first to admit that there's a lot of details I don't know about IP. Maybe the state of IP in general, and full-duplex real-time IP specifically, is so immature that Pandora's Box indeed is opened by a full-duplex, real-time app like VoIP. If so, we have to remember exactly what the IP crowd will say when these things happen:

Stupid idiot newbie users! What do they expect, something that works like we said it would? How naive could they be? What -- they think they're just going to punch a few buttons and talk, fer chrissakes?

stephenpcooke 12/5/2012 | 2:01:01 AM
re: Mobile Skype: Quality Issues? a VoIP company would have to notice that very large numbers of new accounts were being created and that these accounts were all being charged to a small set of credit cards.

implying that a new VoIP company was looking to limit their revenue....? Companies do not care what credit card meets their authorization, just that one does. Once it has been authorized they are good to go. I am sure some VoIP salespeople are not as honest as others and, even if they knew this was about to happen, would enjoy it because they would still be getting paid, though by the insurance companies of the credit cards.

T really do not think that either the modem POTS or the VoIP attack is likely. However the POTS attack would seem to be much easier to pull off.

Come on dljvjbsl, you are much smarter than this!
itisi 12/5/2012 | 2:01:02 AM
re: Mobile Skype: Quality Issues? Be it noted that there's an asymmetry in DOS attacks in the different technologies. On the one hand, it's much easier to launch DOS in the IP space. On the other hand it's much harder to resist in the POTS space.

When I was a kid in a town in the Great Plains with one movie theater, me and my geek buds wanted to see Dr Strangelove (dates me exactly?). But the local theater didn't have it on the schedule. So we launched an early 60's DOS. The half-dozen of us just dialled the theater office a lot, each time asking for the movie.

Finally they said the hell with it, and showed it. I remember seeing it, with just about, oh, a half-dozen folks in the audience.

You don't need a fancy DDOS zombie net to cripple most end points on the POS net.
dljvjbsl 12/5/2012 | 2:01:02 AM
re: Mobile Skype: Quality Issues?
The power of DoS attacks comes from the ability to infact tens of thousands or even millions of computers that can then act at the hackers command or at least their coordination. Long distance phone calls cost money to the owner of the phone line. VoIP calls can be charged to stolen credit cards without the owner being any the wiser, until they get the bill. Audio spam is the same except that the spamming company could legally pick up the bill.


As you are implying here, VoIP networks are networks. A VoIP user will be assigned an incoming proxy and will only be able to use that proxy. He will not have the crytographic information to use another. So the number of calls that can be made from a single user computer is low.

If a worm attempts to create large numbers of accounts with a stolen credit card, it will face the same problem. These accounts must be separate and distinct. So, a VoIP company would have to notice that very large numbers of new accounts were being created and that these accounts were all being charged to a small set of credit cards.

The number of accounts is something that a VoIP company would tend to track very closely. The sales manager would excitedly report to his CEO that we have gained 50 million accounts in the previous month. This would seem to be something that people would notice.

T really do not think that either the modem POTS or the VoIP attack is likely. However the POTS attack would seem to be much easier to pull off.
stephenpcooke 12/5/2012 | 2:01:03 AM
re: Mobile Skype: Quality Issues? Hi,

This has been bothering me for quite some time, since it was first used as a rebuttal against my suggestion that DoS attacks were now possible on the POTS network. Yes, I am slow as mentioned before but I may have some points worthy of discussion. Here are a couple of givens:

- Prior to the advent of VoIP, you actually had to phone another phone number to ring their phone.

- most people who use dial-up Internet services turn their computers off at night or when they are not interested in surfing/doing other computer-related activities.

- when you phone another number, particularly a long distance number, they show up on your phone bill.

OK, so now that we have a few datapoints, lets analyze:

DoS attacks require coordination (ie: all virus/worm-infested computers have to act in unison, or close to it). If the computers are turned off, they can't be used.

DoS attacks via worms/viruses are meant to be generally undetectable by the computer user, at least until the event happens. Things like the phone being dialed when someone else wants to make a call, the phone going off-hook and modem sounds in the middle of the night, etc. would all require some form of investigation and potentially anti-virus-type response.

The amount of logistical coordination required is unimaginable, especially for a hacker who is more interested in writing code than studying human behavioural patterns (eg: what people will download, when they are logged on, when they go to sleep at night, when they work, are there other people at home who might notice the phone being dialed, etc.).

The power of DoS attacks comes from the ability to infact tens of thousands or even millions of computers that can then act at the hackers command or at least their coordination. Long distance phone calls cost money to the owner of the phone line. VoIP calls can be charged to stolen credit cards without the owner being any the wiser, until they get the bill. Audio spam is the same except that the spamming company could legally pick up the bill.

Bottom line: VoIP has enabled DoS attacks where there was no realistic possibility before. It has opened Pandora's box to possibilities, good AND bad, that have never been considered before in the PSTN.
fgoldstein 12/5/2012 | 2:01:28 AM
re: Mobile Skype: Quality Issues? >> The only way to tax VoIP now is when a VoIP call intersects the legacy PSTN.
> Oh, horseshit. Just tax the broadband connection, idiot.

Interstate telecommunications services are currently taxed. DSL, provided wholesale to ISPs by LECs, is taxed. Cable modems are a different animal -- the Ninth says it contains a Telecommunications Service, the FCC does not, and it'll probably end up at the Supremes. (Of course it has telecommunications, but if that's not sold as a service, it's not taxed.)

Most such taxation is based on gross revenues. Flat rate usage plans have no discrete revenues for the calls. Thus nothing to tax. All legal.

PSTN interfaces, though, are chargeable by the PSTN operator. The argument, as I've noted before, is over what rate should be paid by whom to whom when.
technonerd 12/5/2012 | 2:02:10 AM
re: Mobile Skype: Quality Issues? All kidding aside you were out of line with that post.
Oh, get over yourself.
aswath 12/5/2012 | 2:02:25 AM
re: Mobile Skype: Quality Issues? Apropos the discussion in this thread, there is a news item that people are using IP Relay Service to scam retailers. Start with http://slashdot.org/articles/0... and follow the news story for additional details.
technoboy 12/5/2012 | 2:02:28 AM
re: Mobile Skype: Quality Issues? All kidding aside you were out of line with that post.
dljvjbsl 12/5/2012 | 2:02:28 AM
re: Mobile Skype: Quality Issues?
How about: "Sir, I beg your pardon. I am not shameless!"


From your postings it is obvious that you have only a limited understanding of telephony and next to no understanding of IP.
technonerd 12/5/2012 | 2:02:33 AM
re: Mobile Skype: Quality Issues? How do you reply to someone who calls you a shameless whore?
How about: "Sir, I beg your pardon. I am not shameless!"
DanJones 12/5/2012 | 2:02:33 AM
re: Mobile Skype: Quality Issues? Similar problems with WiMax as WiFi as far I can see.

- No voice QoS in standard, so no way of prioritizing voice traffic on a network, although this might change with another standards revision I suppose.

- No handoff between WiMax base stations at least in 802.16a revision d spec that is being used for initial chipsets, less of a problem than with WiFi (as WiMax has more range) but still hardly "mobile" communications. The .16a rev e spec might fix this?

- First true WiMax chipsets are not even on the market yet. Remember how long it took to get WiFi as standard issue on laptop PCs? Four odd years from when the standard first hit. 802.11 is certainly not standard issue in handsets yet. Why should WiMax be adopted any faster?

Just some thoughts

Dan Jones
Site Editor, Unstrung
eman 12/5/2012 | 2:02:36 AM
re: Mobile Skype: Quality Issues? One of the threads in this discussion talked about the possibility of using 802.11 WiFi networks to deliver voice service (as potential competition to cellular). It seems there are technical limitations to WiFi (i.e. few, if any, handoff protocols supported, limited range per WiFi site, many sites needed, etc.).

What about the networks based on 802.16 WiMAX? Since 802.16 networks are designed to cover a wider footprint, do they have the capability to be viewed as potential competition in the mobile voice and broadband space?
technoboy 12/5/2012 | 2:02:38 AM
re: Mobile Skype: Quality Issues? regarding the FCC's ruling on Febuary 12th of 2004, the same day they also started a working group to further explore what their role should be in the future regulatory role on VOIP or IP communications. This working group is looking at all aspects of IP communications and they specifically state that this will continue to evolve and change over time. This is essentially what I stated. It is my contention that eventually IP communications will be regulated and taxed appropriately. Time will tell.
technoboy 12/5/2012 | 2:02:39 AM
re: Mobile Skype: Quality Issues? Hi Dreamer101

You really got me on that one!!!!

Did that take you all night???

technoboy 12/5/2012 | 2:02:39 AM
re: Mobile Skype: Quality Issues? To fkittred:

If you are talking about someone who decides to install their own VOIP client or use windows messenger to communicate with someone else that is very true. The message thread was in the context of why Carriers were moving to VOIP or IP communications and away from the PSTN. Some believe it is only to remove themselves from regulation. I believe there are several reasons as to why they are moving to this technology but I also believe they will be regulated on some level and that regulation will probably expand to cable operators and potentially others who would offer services as well. In this context it would not be necessary to monitor and measure usage. As a customer of these services I would likely be charged a flat rate for basic service and additional charges for additional features. On top of those charges, the carrier could pass down fees that the regulatory agency decided to charge. Where monitoring of RTP media streams comes into play is when things like CALEA are required.
technoboy 12/5/2012 | 2:02:40 AM
re: Mobile Skype: Quality Issues? technonerd wrote:... including this VoIP for which you have become such a shameless whore. 'Boy, I really hope you have a buttload of options, because no one should just give away his integrity and intelligence. At the very least, you should do what they do in Silicon Valley: sell it to the highest bidder.

How do you reply to someone who calls you a shameless whore????????
dljvjbsl 12/5/2012 | 2:02:40 AM
re: Mobile Skype: Quality Issues?
How much you wanna bet the IP-heads here will say it's the user's fault?


10 cents and a bottle of Pepsi
dljvjbsl 12/5/2012 | 2:02:41 AM
re: Mobile Skype: Quality Issues?
Gee, I think I hit a nerve.


There you go again
technonerd 12/5/2012 | 2:02:42 AM
re: Mobile Skype: Quality Issues? Check this out:

Bluetooth May Put You
At Risk of Getting 'Snarfed'

By JEREMY WAGSTAFF
Staff Reporter of THE WALL STREET JOURNAL
April 15, 2004; Page D3

If you spot someone tailgating you on the road or standing next to you wearing a backpack, then watch out: You may have been "snarfed." All the data on your cellphone, including addresses, calendars, whom you called and who called you, may now be in that person's computer.

Many cellphones use Bluetooth technology, which allows them to communicate wirelessly with other Bluetooth-equipped devices -- computers, personal-digital assistants and other cellphones. This means you don't need a cable, for example, to synchronize the address books on your laptop and your cellphone. It is convenient, but that makes it possible for someone to steal your data, or even hijack your cellphone for their own purposes.

Last year, London security consultant AL Digital spotted flaws in the way some Bluetooth cellphones swapped data with one another -- flaws that could be used to gain unauthorized access to everything stored on that phone without the user ever knowing. AL Digital's Adam Laurie, who discovered the problem, shared his findings with cellphone makers and with the public (leaving out the detail that might allow ne'er-do-wells to copy his experiments at street level). He termed the trick Bluesnarfing.

Not a lot has happened since then. Nokia Corp., the market leader in the cellphone industry, acknowledges the flaw but says in an e-mail response to questions that it is "not aware of any attacks against Bluetooth-enabled phones." Sony Ericsson, a joint venture of Telefon AB L.M. Ericsson and Sony Corp., didn't reply to an e-mail. Even those highlighting the danger say they haven't heard of specific attacks.

Still, these attacks -- also known as Bluejacking -- nevertheless are possible. Mr. Laurie cites a scenario in which paparazzi could steal celebrity data. He says he was able, with permission, to snarf from a friend's phone details of her company's shops, door codes and safe combinations. "There's any number of angles you can look at, and they are all bad as far as I can see," he says.


O.K., it's not a VoIP story, but these Bluetooth cellphones are minature computers. And they're not secure. And we can see how much the vendor cares. How much you wanna bet the IP-heads here will say it's the user's fault?
technonerd 12/5/2012 | 2:02:42 AM
re: Mobile Skype: Quality Issues? The PSTN had plenty of time to mature prior to it being regulated.
Among other things, the IP propellerheads are too arrogant to study arcane subjects like "history." Sort of like the Bush administration, which apparently is unaware that the Brits entered Baghdad in 1920 proclaiming themselves "liberators" only to face insurrection and upprising within months.

"History" will show that AT&T won its monopoly status in return for accepting government regulation. All of this happened prior to 1920. But hey, IP liberates us from really having to know anything, right technoboy?


I know plenty about what is going with the regulatory issues
You sure could have fooled me.


and you should know that everything to date is still a work in progress.
... including this VoIP for which you have become such a shameless whore. 'Boy, I really hope you have a buttload of options, because no one should just give away his integrity and intelligence. At the very least, you should do what they do in Silicon Valley: sell it to the highest bidder.


Witness the intention of the californina PUC and Illinois PUC to press the issue for regulation. The FCC has also been notified by the justice department that exempting VOIP from some of these regulatory requirements such as CALEA may violate sections of the patriot act. So this matter is far from settled.
The FCC has ruled VoIP to be an application on a deregulated network. This is the result of a corrupt arrangement with the RBOCs, their suppliers and the major VCs in California.


The industry is doing the right thing at the moment which is attempting to work with the various agencies to meet their specific requirements for E-911 and CALEA.
"The industry" never does the "right thing." It acts solely in its economic self-interest. Nothing else matters.
technonerd 12/5/2012 | 2:02:43 AM
re: Mobile Skype: Quality Issues? The only way to tax VoIP now is when a VoIP call intersects the legacy PSTN.
Oh, horseshit. Just tax the broadband connection, idiot.
technonerd 12/5/2012 | 2:02:44 AM
re: Mobile Skype: Quality Issues? There you go again
And again.
technonerd 12/5/2012 | 2:02:44 AM
re: Mobile Skype: Quality Issues? Keep up the commentary. You continue to demonstrate your ignorance. You wouldnt know the truth if it bit you in the ass. What a condesceding littel twit you have become.
Gee, I think I hit a nerve.
dljvjbsl 12/5/2012 | 2:02:44 AM
re: Mobile Skype: Quality Issues?
It's possible, I suppose, to build a carrier-class PBX installation, even using stock systems, but it's not the norm. Two hours/year downtime is viewed as the norm. That's horrible for a CO. It's fantastic for an IP network.


I take your point about the reliability of some enterprise systems. Some customers are indifferent to power failures and will run the system non-redundant in both power and control. Other customers such as police departments, hospitals etc are not indifferent and will pay for the redundancy they require. The Metropolitan police in London, for example, runs a network of PBXs.

sgan201 12/5/2012 | 2:02:44 AM
re: Mobile Skype: Quality Issues? Hi Technoboy,
In summary, what you are trying to say is VoIP is not matured enough to replace POTS yet..

"The truth will set you free!!"

Dreamer
fkittred 12/5/2012 | 2:02:44 AM
re: Mobile Skype: Quality Issues? I don't think you understand the technical issues involved in taxing VoIP, nor the rapidity with which the situation is evolving.

One can't tax what one can't measure. There is no way to detect
or measure encrypted VoIP streams between two consenting adults. There is no reason the coordination functions of setting up a VoIP "call" can't be peer-to-peer, like Napster, or off-shore.

The only way to tax VoIP now is when a VoIP call intersects the legacy PSTN. However, as time goes on, more and more pure VoIP calls will be made which do not touch the PSTN. In an article in lightreading today there was a description of the cable companies' move to set up an ENUM system so that they can connect calls directly between each other without using the PSTN as an intermediary.

If you only tax the legacy network, then it will die even quicker. Please see last week's news stories on the S&P credit watches placed on VZ, SBC, Qwest, etc.

regards,
fletcher
technoboy 12/5/2012 | 2:02:45 AM
re: Mobile Skype: Quality Issues? Reply to Fgoldstein: If you have worked in the telecom regulation for so long than you know that things change and in this case VOIP will not be an exception. If you read my post I am not trying to have it both ways. The PSTN had plenty of time to mature prior to it being regulated. I know plenty about what is going with the regulatory issues and you should know that everything to date is still a work in progress. Witness the intention of the californina PUC and Illinois PUC to press the issue for regulation. The FCC has also been notified by the justice department that exempting VOIP from some of these regulatory requirements such as CALEA may violate sections of the patriot act. So this matter is far from settled. The industry is doing the right thing at the moment which is attempting to work with the various agencies to meet their specific requirements for E-911 and CALEA.
fgoldstein 12/5/2012 | 2:02:45 AM
re: Mobile Skype: Quality Issues? > Enterprise systems have as stringent reliability requirements as the PSTN.

Well, no. I've been a telecom manager, and a consultant to enterprise telecom networks, and I know PBXs quite well. They're nice, but not built to CO standards.

For one thing, many PBXs are powered by AC, not DC. Some are touchy, although not as bad as they used to be. I cut over a Rolm CBX in 1979, a feature-rich digital switch. It had a touchy power supply that would reboot the system, dumping calls, when it noted a half-cycle (!) of AC power slippage. Which was often. Quick, onto the big (roomful of batteries) UPS! And it needed 30% humidity, so there was a rotating-drum humidifier in the room, along with the Edpac 5 ton ACs. Okay, that was an extreme example. Maybe.

But just a couple of years ago, I watched a Nortel Meridian SL-1 Option 71 (really, a good PBX as they go) take a day-long dive. Why? There was a glitch in the fiber feeding three Sprint channelized DS-1s. That caused 72 simultaneous E&M-lead actions, which overloaded the CPU, and it couldn't reset itself. And while it was DC powered, a bad electrical storm brought the whole thing down for a day when a lightning strike induced power into the station lines. Lots of emergency parts replacement. No, CO switches don't do this.

The big hosiptals here in Boston are on Centrex, on ancient 5ESS big iron. Reliable though. It's possible, I suppose, to build a carrier-class PBX installation, even using stock systems, but it's not the norm. Two hours/year downtime is viewed as the norm. That's horrible for a CO. It's fantastic for an IP network.
fgoldstein 12/5/2012 | 2:02:46 AM
re: Mobile Skype: Quality Issues? technoboy > For someone so wise on telecom regulation, you are still making a rather speculative assumption. VOIP will be regulated if not at the federal level than certainly at the state level. There are literally billions ($$$)of reasons why this will happen. I will also predict that the MSO's will be regulated at some point as well. All that is happening in the short term is that regulators are providing a kind of grace period so that technology can grow and take hold. Once it is mature enough, the government will come looking for its piece of the pie. I would think you would realize this.

Wow, 'boy, you really don't know what's going on in the regulatory world, do you?

In the case of non-PSTN VoIP to VoIP, the FCC has ruled that it's purely a federal matter, not subject to state jurisdiction. In the case of PSTN-to-computer VoIP, it's tentatively a federal matter; there's an open docket (go to FCC ECFS for 04-36 to see the record) on how to deal with VoIP more permanently. The current FCC believes (for no particularly valid legal reason, but this is the Bush administration we're talking about, so law is only a suggestion) that the Telecom Act only applies to networks and components that were in place and operating before the Telecom Act was passed in 1996. So while there is no legal reason for VoIP to exempt the ILECs from dominant-carrier regulation, Powell would certainly use it as an excuse to do so. And the MSOs are governed by the Cable Act of 1992, which eviscerated the bulk of state or local regulatory authority.

I work in telecom regulation, and deal with the subtleties on a daily basis, including the fact that different states view their power differently, so there's no consitency. But there are real limits on what a state can do to constrain an ILEC. Monopolists currently have a certain droit de seigneur.

Regulators should not give VoIP a bye. If it is really so great, it needs no favoritism. If it needs favoritism, it shouldn't succeed. You can't have it both ways.
dljvjbsl 12/5/2012 | 2:02:48 AM
re: Mobile Skype: Quality Issues?
Why would the malicious party send the requests to one proxy? A handful of zombies can screw up lots of proxies at once.


The VoIP network is a network. The zombie would be assigned to a proxy. It would not have the cryptographic knowledge to conenct to another proxy
fgoldstein 12/5/2012 | 2:02:48 AM
re: Mobile Skype: Quality Issues? me >> one Fast Ethernet interface on a cheap server can generate thousands of call VoIP call requests per second.

dljvjbsl > Which will be ignored by the proxy to which they are sent

Why would the malicious party send the requests to one proxy? A handful of zombies can screw up lots of proxies at once.
stephenpcooke 12/5/2012 | 2:02:49 AM
re: Mobile Skype: Quality Issues? I would submit that in your scenario the PSTN attack would be serious and the VoIP attack would be trivial.

How many attacks such as this would it take to damage the credibility of carriers with their customer base? On the economic front, how many would it take before the carrier(s) were forced to disconnect their POTS & VoIP networks? Given that the impact would be serious on the PSTN, which is regulated by the FCC, how many such attacks would it take before legislative action were taken? If legislative action were taken what would the impact be on the business case of VoIP providers (including cable companies)?

There are no easy answers here that I can see. Obviously this issue goes well beyond the capabilities of IP or any other newer technology.
technoboy 12/5/2012 | 2:02:49 AM
re: Mobile Skype: Quality Issues? technonerd:I have a suggestion. Try a tactic unknown among IP evanglists: Tell us the truth before we drag it out of you.


I have a suggestion for you as well. Keep up the commentary. You continue to demonstrate your ignorance. You wouldnt know the truth if it bit you in the ass. What a condesceding littel twit you have become. FYI, why dont you answer the quetions posed to you by the other posters in the thread. Now go whip out your next snappy retort Technonerd. I know you cant wait.
dljvjbsl 12/5/2012 | 2:02:50 AM
re: Mobile Skype: Quality Issues?
Enterprise voice systems are glorified intercoms. They can't be compated with the PSTN


There you go again
technonerd 12/5/2012 | 2:02:50 AM
re: Mobile Skype: Quality Issues? Sorry, I should have put an emoticon there when I said 'berated' ...but you are right, admitting this will most probably limit my stock options in the future.

Wow, do people's bosses actually read these forums? If so, then they not we, are the ones with way too much time on their hands!!
technonerd 12/5/2012 | 2:02:50 AM
re: Mobile Skype: Quality Issues? Enterprise systems have as stringent reliability requirements as the PSTN. In the case of emergency services such as a police network, the implications of a failure would be far more serious than te lsoss of serive to a pizza parlor with a PSTN number.

Enterprise voice systems are glorified intercoms. They can't be compated with the PSTN.
stephenpcooke 12/5/2012 | 2:02:50 AM
re: Mobile Skype: Quality Issues? Technonerd,

Sorry, I should have put an emoticon there when I said 'berated' ...but you are right, admitting this will most probably limit my stock options in the future. ;)
technonerd 12/5/2012 | 2:02:51 AM
re: Mobile Skype: Quality Issues? Worms are not affected by firewalls. They use subtrefuge to trick users into installing them on their computers.

Sorry. Let me rephrase. A lot of people have elementary firewalls and virus protection ...
dljvjbsl 12/5/2012 | 2:02:51 AM
re: Mobile Skype: Quality Issues?
These are matters of opinion. I stand by my statements. Telephones and the networks that connect them are towers of reliability, usability and functionality compared to computers and the fragile IP networks that connect them.


So the enterprise networks within hospitals and police stations are not as reliable as a class 5 CO. I think people would be very very surprised about this.

Are you gong to admit that you made a mistake?
technonerd 12/5/2012 | 2:02:51 AM
re: Mobile Skype: Quality Issues? What evidence do you have that enterprise systems are any less reliable than PSTN systems?

I've never been connected to an IP network, enterprise or public, that didn't go down at least once every few months. By contrast, I can count on the fingers of one hand (which most of them unused) the occasions in the past 25 years when I've been unable to make a landline phone call.

Cellular networks are another story; there, we trade blocking and lower transmission quality for mobility. But even they have been steadily improving, and are now far more reliable than IP networks.
technonerd 12/5/2012 | 2:02:51 AM
re: Mobile Skype: Quality Issues? I'm happy to engage in an objective debate on the merits and negatives of Voice Over IP technology. I will probably regret saying this but there are negatives.
I have a suggestion. Try a tactic unknown among IP evanglists: Tell us the truth before we drag it out of you.


The issue I have had with the debate is that it is not objective. I ask people to make objective assessments and not take shots.
Who says that you can't "take shots" while being objective?


There are egos involved (mine included) and it makes it difficult to elevate the level of discourse.
"Elevating the level of discourse" typically means not questioning the mushy millennialism pumped out from the IP in roughly the same quantities as sewage from Manhattan. Think of me as one of those towns down South who wouldn't allow the New York City garbage barge to dock in their port.


I would also ask for people to have some perspective on where VOIP is in the evolutionary cycle as opposed to where the PSTN is. There are statements made based on assumptions that are inaccurate. There is a notion that somehow we will flick the switch and cut over to an IP based network. That is not the case. It is a migration that will take place over time. Lots of issues to resolve but I am confident they will be resolved. There are alot of smart people working on it.
There were "a lot of smart people" working on the Internet frauds, too. Guess what? It didn't make them any less fraudulent.
dljvjbsl 12/5/2012 | 2:02:51 AM
re: Mobile Skype: Quality Issues?
I've never been connected to an IP network, enterprise or public, that didn't go down at least once every few months


You can wiggle all you want. You made statments about enterprise voice systems that are clearly incorrect.

Enterprise systems have as stringent reliability requirements as the PSTN. In the case of emergency services such as a police network, the implications of a failure would be far more serious than te lsoss of serive to a pizza parlor with a PSTN number.

technonerd 12/5/2012 | 2:02:52 AM
re: Mobile Skype: Quality Issues? The first one was when you stated that enterprise applications had less stringent reliability requirements than PSTN. The second one was when you said that enterprise equipment was less reliable than PSTN equipment.

These are matters of opinion. I stand by my statements. Telephones and the networks that connect them are towers of reliability, usability and functionality compared to computers and the fragile IP networks that connect them.
dljvjbsl 12/5/2012 | 2:02:53 AM
re: Mobile Skype: Quality Issues? ,b>
This was once more plausible than it now is. More and more individual computers have the rudimentary firewalls that would prevent this sort of crude attack


Do you know anything?

Worms are not affected by firewalls. They use subtrefuge to trick users into installing them on their computers.

Are you going to admit this mistake?
technoboy 12/5/2012 | 2:02:53 AM
re: Mobile Skype: Quality Issues? Stephencooke,

I'm happy to engage in an objective debate on the merits and negatives of Voice Over IP technology. I will probably regret saying this but there are negatives. The issue I have had with the debate is that it is not objective. I ask people to make objective assessments and not take shots. There are egos involved ( mine included) and it makes it difficult to elevate the level of discourse. I would also ask for people to have some perspective on where VOIP is in the evolutionary cycle as opposed to where the PSTN is. There are statements made based on assumptions that are inaccurate. There is a notion that somehow we will flick the switch and cut over to an IP based network. That is not the case. It is a migration that will take place over time. Lots of issues to resolve but I am confident they will be resolved. There are alot of smart people working on it.
technonerd 12/5/2012 | 2:02:54 AM
re: Mobile Skype: Quality Issues? I have a problem with your desire to sit there and spin nonsense so you can look at your statements and pat yourself on the back on how your smart ass remarks. It overshadows the valid points you do make.

You're really not at your best when you're whining.
dljvjbsl 12/5/2012 | 2:02:55 AM
re: Mobile Skype: Quality Issues?
I will admit a mistake if I make it


You have made mistakes every time you mention enterprise networks.

The last one was actually two mistakes in one. The first one was when you stated that enterprise applications had less stringent reliability requirements than PSTN. The second one was when you said that enterprise equipment was less reliable than PSTN equipment.

Are you going to admit these or jsut accuse people of being stupid and corrupt.
technonerd 12/5/2012 | 2:02:56 AM
re: Mobile Skype: Quality Issues? In any event, it is a scenario that is not particular to VoIP. There are millions of computers with dial up modems attached. A worm which takes posession of vast numebers of computers as you say then could make large numbers of calls simultaneously towards a specific class 5. This would exhaust the trunks to the class 5 and isolate it.

This was once more plausible than it now is. More and more individual computers have the rudimentary firewalls that would prevent this sort of crude attack. Mass-market VoIP would inevitably be more vulnerable to disruption than PCM networks.
technonerd 12/5/2012 | 2:02:56 AM
re: Mobile Skype: Quality Issues? For someone so wise on telecom regulation, you are still making a rather speculative assumption. VOIP will be regulated if not at the federal level than certainly at the state level. There are literally billions ($$$)of reasons why this will happen. I will also predict that the MSO's will be regulated at some point as well. All that is happening in the short term is that regulators are providing a kind of grace period so that technology can grow and take hold. Once it is mature enough, the government will come looking for its piece of the pie. I would think you would realize this.

The government always wants its pound of flesh, but the FCC's stance on VoIP goes well beyond encouraging the growth of an infant application. The FCC shows every sign of wanting to allow the Regional Bells to use VoIP to kill the Telecom Act. This is part of a corrupt deal between the RBOCs, their suppliers (in many cases coerced), and Republican officeholders in need of campaign contributions.

It's pretty well known in Washington, D.C. that Powell, the chairman of the FCC, has harbored ambitions to run as a Republican for the U.S. Senate from Virginia. This campaign will be funded by the Bells and their vendors in return for services rendered. Just wait.
dljvjbsl 12/5/2012 | 2:02:57 AM
re: Mobile Skype: Quality Issues?
By multiple this might imply that the credit card information is distributed in a worm that creates an account with a "trusted" VoIP service supplier


Waht you are talking about is a computer generated mass calling event. We have discussed this before.

In any event, it is a scenario that is not particular to VoIP. There are millions of computers with dial up modems attached. A worm which takes posession of vast numebers of computers as you say then could make large numbers of calls simultaneously towards a specific class 5. This would exhaust the trunks to the class 5 and isolate it.

In case of a VoIP attack only the trunks between the class 5 and the VoIP gateway would be affected and the class 5 would carry on as before.

I would submit that in your scenario the PSTN attack would be serious and the VoIP attack would be trivial.

technonerd 12/5/2012 | 2:02:57 AM
re: Mobile Skype: Quality Issues? As I have been berated by technonerd for admitting, I am no genius but I am sure there are some nasties out there that can make this sort of thing happen in any way that suits their purpose.

I doubt I've berated you, Steve, and I can't imagine berating anyone for admitting they're not a genius. I usually berate people for insisting that they are geniuses as the propellerhead IP evangelists are wont to do.

That said, if I have berated you then I take it back and apologize. Unlike our president, I will admit a mistake if I make it. I think your posts have been reasoned and intelligent.
stephenpcooke 12/5/2012 | 2:02:58 AM
re: Mobile Skype: Quality Issues? Lets add in going after multiple VoIP providers in multiple countries as well...

I am sure you are getting the idea. Let me also caution, this is only a guess as to what is possible. Once someone figures out how to do these things there will be hundreds of them because this is what hackers do. How long will it take to kill every VoIP provider because they have been cut off from access to the PSTN as a purely defensive measure on the part of the PSTN providers? How much damage will have been done before that happens?

Smarter people than me will figure out much more elaborate and difficult to detect and stop schemes just because they can. If they are pissed at someone, they are probably more dedicated. Where does it stop?
stephenpcooke 12/5/2012 | 2:02:58 AM
re: Mobile Skype: Quality Issues? Here are some followup questions to consider:
Given that access to the PSTN network via VoIP technology IS sold, put yourself in the position of a hacker or spammer. What do you do?

Some possibilities might include stealing some credit cards and opening multiple 'valid' accounts. By multiple this might imply that the credit card information is distributed in a worm that creates an account with a "trusted" VoIP service supplier. UserID's and passwords are generated automatically by the service provider on verification of the credit card. This information is then made accessible to the worm via some website perhaps along with a list of numbers to call. At the designated time the worm begins its work of logging into the provider and calling the numbers with or without an associated message, depending on the goal (perhaps spam would have a pointed message, network disruption would involve connection attempts and hang ups, etc.).

In the case of audio spam, no worm is required but multiple accounts can be created for this exact purpose within some corporate environment. Multiple corporate credit cards could be used to set up these accounts.

I could go on but this doesn't seem terribly difficult to me. How is this different from other calling schemes? It is a matter of scale and time. Telemarketers generally run off a PBX which is circuit-oriented. The PBX can only support a given number of circuits at any one time which was taken into account by the network when they ordered service. There are generally people who want to speak with other people so there are time constraints as well. Neither of which apply to the case I have described. The only limitations I can see are the number of active clients on any computer at any one time and the number of computers being used at a time.

As I have been berated by technonerd for admitting, I am no genius but I am sure there are some nasties out there that can make this sort of thing happen in any way that suits their purpose.
stephenpcooke 12/5/2012 | 2:02:58 AM
re: Mobile Skype: Quality Issues? I think the more important question was the second one:

"Is access to these networks openly sold as a business opportunity?"
technoboy 12/5/2012 | 2:02:58 AM
re: Mobile Skype: Quality Issues? Technoboy wrote: Us propellerheads who have been in the business for a long time are quite aware of the technical issues surrounding VOIP.
Actually, it's "we propellerheads," but I digress. If you're so aware of the technical issues, then why all the goddamn evangelism for a science project still in progress? technonerd responded: My "obsession" is with bursting the hype bubbles that have allowed so many California shysters to team up with so many New York shysters to steal so much money.

Technonerd when did I ever in any of my posts claim this was the greatest technology or over hype it. Im not a sales rep or marketing rep but I am an engineer and I have been working with service provider and enterprise customers in this space for a long time. So I would put my knowledge of how their networks are evolving up against your knowledge anytime. In fact in several posts I agreed that there has been a lot of hype on the technology. I have also stated there will be a co-existence of the PSTN and VOIP for some time to come. No one ever said we would flash cut the PSTN OVER TO VOIP. It is pretty obvious that you are up to your old ways of just trying to piss people off. Especially with the reply to my statement regarding the work that still needs to be done. Too bad you are not able to interject a little objectivity into your point of view. I have a problem with your desire to sit there and spin nonsense so you can look at your statements and pat yourself on the back on how your smart ass remarks. It overshadows the valid points you do make. Since I doubt you have had any real interaction with IP voice technology in the last year I submit you are not qualified to judge its functinality or viability in production networks today. Rather than sit back and crap on everyone, why dont you go educate yourself on where the technology is today and then make an objective judgement.

For someone so wise on telecom regulation, you are still making a rather speculative assumption. VOIP will be regulated if not at the federal level than certainly at the state level. There are literally billions ($$$)of reasons why this will happen. I will also predict that the MSO's will be regulated at some point as well. All that is happening in the short term is that regulators are providing a kind of grace period so that technology can grow and take hold. Once it is mature enough, the government will come looking for its piece of the pie. I would think you would realize this.
dljvjbsl 12/5/2012 | 2:02:59 AM
re: Mobile Skype: Quality Issues?
How many of these are easily accessible via the open Internet?


The issue is that VoIP protocols are designed so that their components are not open to anyone on the Internet. The VoIP network is just that. It is a network of trusted elements. They will accept traffic only from known and trusted sites.

As to the issue of the open network, I still ask how many hackers have penetrated the security around banking websites?


I can recall the great concrn that people expressed about enterprise security and industrial espionage. This was put into perspective for me once by a speaker at a conference. He asked whether a hacker would want access to the CEO's strategy PowerPoint slides or accounts payable. The types of systems that enterprises would want became clearer to me that day.
stephenpcooke 12/5/2012 | 2:02:59 AM
re: Mobile Skype: Quality Issues? dljvjbsl wrote:

Some of the enterprise networks that use this technology are:

a) police departments
b) hospitals
c) E911 systems
d) naval warships
e) 24 hour call centers


How many of these are easily accessible via the open Internet? Is access to any of these networks openly sold as a business opportunity?
dljvjbsl 12/5/2012 | 2:03:00 AM
re: Mobile Skype: Quality Issues?
Most IP vendors are primarily selling their stuff into enterprise networks, Cisco of course being the prime example. An enterprise network can easily be shut down for a couple hours on Friday night. A carrier network can't. Regulators won't allow it, nor will customers. If the carrier network is down in the middle of the night and someone dies because they couldn't call the ambulance, it will be on the front page the following day and the head of the Public Service Commission will be on the phones that now work asking for some explanations right away. And when the next rate hearing comes up, memories will be long.


Some of the enterprise networks that use this technology are:

a) police departments
b) hospitals
c) E911 systems
d) naval warships
e) 24 hour call centers
f) ...

Enterprise systems are used in life and mission critical situations. Failures are simply not acceptable in these systems.

What evidence do you have that enterprise systems are any less reliable than PSTN systems?

With some evidence in that regard, I may believe that you have a little understanding of the capability and functionality of PBXs.

As to visting the telephone plant, I ahve done that several times in my career. I have been involved in the design of two CO platforms (mulitple COs were developed from these). You have undoubtedly used my circuits. They will be everywhere in your beloved PSTN. Do not assume that you are the only one with experience in this area. You are not.
stephenpcooke 12/5/2012 | 2:03:02 AM
re: Mobile Skype: Quality Issues? Have the people on these boards stopped and considered the levels of experience on both sides of this debate? What does this tell us?

It tells me that there are some seriously experienced people out there who believe in the approach of convergence. I respect this, I really do. I also have been in the technical trenches as the system design authority of the optical transport platform that was #1 in the world at that time. I thought that I was pretty good too. I was painfully forced to acknowledge that there were some drastic improvements that could have been made, both in my product and its underlying technology, some of which had been simple design trade-offs earlier in the process that had simply not been recognized as providing the necessary added value.

I respect the knowledge and dedication of the people who are advocating VoIP technology but, on occassion, I have to question the closed mindedness of accepting that something WILL have been missed that will cause serious problems. Those of us who have been around for more than a few years will have to admit that introduction of technology is never a smooth process in the best of conditions. Prior to the advent of VoIP technology the PSTN was an isolated network (ie: it was exceedingly difficult for external malicious entities to do drastic damage on any large scale). VoIP has changed that.

How many viruses have been developed in the last 10 years? How many worms? 10 years ago did any of these exist? How about spyware and adware? We have had to learn about firewalls and anti-virus programs but, lets face it, these were pretty much side issues because no one's life depended on the Internet.

Did Microsoft intentionally design in the OS holes utilized by viruses? I am sure their margins would be much better if they didn't have to keep releasing patches to close them. I'm sure technoboy's SIP proxy works fine but is it bulletproof? Is it the weakest link in the VoIP story?

I have been a test engineer in my career as well. I have said before that, even when I was with a third-party test organization, I saw no totally standards compliant equipment. This generally leads to trouble tickets or problem reports, etc. that have to be tracked. How many products have been released with open, serious and documented problems, let alone the undocumented ones? Answer: Every single one of them.

I have presented some examples that I felt were areas of concern that I personally have not heard being addressed. Perhaps some of them have been, but unless God himself is designing the equipment and the network, there will be serious issues; to deny this is nothing but arrogance. What many of us have tried to say, technonerd most eloquently of course, is that by combining a relatively bulletproof, isolated network with the wide open, virus-ridden Internet, means that people's lives will be unsuspectedly put at serious risk. To deny this is to throw out the benefits of any experience that you have gained in this industry. Saying that "it will be better this time" has been used with every other technology introduction in history, none have gone smoothly, especially with such a complicated mesh of legacy telecom and datacom equipment that must be supported in any resultant architecture.

"We are taking the next step in technology...", yes, but have we asked the question "Is this the right step"? Has a decent risk analysis been done by non-biased people? What will happen if the FCC, Whitehouse, or other government organization is hit with a successful (however it gets through) DoS attack? Saying it won't get through because of the "trusted" nature of some feature that hasn't yet been implimented is not a sufficient response. Smart hackers are relentless and will learn from the "successes" and failures of themselves and their bretheran. They will adapt much faster than we can design boxes and software patches to stop them. These are things that the PSTN has never had to deal with before and is an ENTIRELY different business model for carriers and equipment providers. It may come down to needing forklifts in every CO in the event that a resident piece of equipment has some previously unbeknownst problem that some hacker in (insert any country in the world here) has figured out how to exploit. When people depend on the communications system as a whole for their lives the stakes are very much different than the current understanding of the Internet or its underlying technology.
lollapalooka 12/5/2012 | 2:03:03 AM
re: Mobile Skype: Quality Issues? Folks,
Watching the electronics industry and the electronics plumbing (our beloved networking/telecom) industry for the past 25 years, it occurs to me that IP vendors follow what I call the Microsoft model of business, where PSTN vendors work to get on approved vendor lists. IP vendors will cobble something together, throw it out there, and if it sticks, start refining it.

I've worked for both types of companies. I feel no stake in the IP vs. PSTN wars, though philosophically I've had to unlearn my love of how effeciently you can pack data into narrowband.

I also don't see IP vendors as evil or liars because they come from an environment where standardized testing is late to evolve. They're now careening towards a highly regulated environment though, and are trying to take the cash cow (IP PBX and IP Centrex) services where they can.

I worked for a great company that made tons of money from selling statmuxes that leveraged the PSTN, but IP networking's "throw bandwidth at it" mentality essentially drove us out of business. But that doesn't mean it's bad. It just means it's different.
Lollapalooka
technonerd 12/5/2012 | 2:03:03 AM
re: Mobile Skype: Quality Issues? I also don't see IP vendors as evil or liars because they come from an environment where standardized testing is late to evolve. They're now careening towards a highly regulated environment though, and are trying to take the cash cow (IP PBX and IP Centrex) services where they can.

In case I have left the impression that somehow I think vendors of PSTN equipment are more honest than vendors of IP equipment, let me un-leave that impression right now. All vendors will tell any lie they can get away with. So will carriers.

However, in practical reality, I think PSTN vendors wind up being effectively forced to be more honest, if we define honesty as making equipment that does what they say it will do. This honesty is forced by a number of factors, including:

- PSTN technology is more mature, which leaves a low-quality vendor with many fewer places to hide if the shit doesn't work

- PSTN carriers are large, powerful buyers who thoroughly study the equipment. They ruthlessly drive cost out while enforcing very high reliability and quality standards, which in turn are functions of regulators and the Bell Labs legacy.

Most IP vendors are primarily selling their stuff into enterprise networks, Cisco of course being the prime example. An enterprise network can easily be shut down for a couple hours on Friday night. A carrier network can't. Regulators won't allow it, nor will customers. If the carrier network is down in the middle of the night and someone dies because they couldn't call the ambulance, it will be on the front page the following day and the head of the Public Service Commission will be on the phones that now work asking for some explanations right away. And when the next rate hearing comes up, memories will be long.

An enterprise will do a complete forklift upgrade; a carrier network is costed out for long depreciation, which means that changes must be incremental. If people lose e-mails on the enterprise network, they can usually be restored. If you can't hold a conversation on the carrier network, you've lost something that can never be restored.

It really boils down to checks and balances. In the PSTN world, a number of forces combine to insure that vendors make equipment that works. None of those things can keep, say, Lucent's finance department from cooking their books. But when Lucent ships a PSTN product to SBC, chances are that piece of iron is going to work.

Hey, propellerheads, I recommend a trip into the telephone plant sometime. Ask one of those idiot technicians (come on, 'fess up, that's what you think of them) to open up a pedestal box for you. The rat's nest inside is a thing of beauty. Want to know why? Because if you look closely, you'll see shit in there that hasn't been touched for 20 years and is still blinking away, carrying calls to phones that work first time, every time.

It's why the ambulance comes in the middle of the night. I suggest we all take some care before we rip that stuff out.
dljvjbsl 12/5/2012 | 2:03:04 AM
re: Mobile Skype: Quality Issues?
You need to read more carefully. I did not say that the carriers have no one trained to run IP networks. What I said is that, if they shitcan the PCM networks and replace them with VoIP networks, they'll have a massive retraining job on their hands.


This is exactly what you said:

Just look on this board and see how untrainable an engineer is; now put yourself in the RBOC position and imagine training everyone on a whole new IP-based network without completely losing reliability and maybe even service at all.


The meaning of these words is clear. The carriers have no one trained in IP since they will have to train everyone and the carriers have no IP networks since the IP-based network will be a whole new network.

Lewis Carrol's Red Queen comes to mind. Words to her mean what she means them to mean and no more.
dljvjbsl 12/5/2012 | 2:03:05 AM
re: Mobile Skype: Quality Issues?
You need to read more carefully. I did not say that the carriers have no one trained to run IP networks. What I said is that, if they shitcan the PCM networks and replace them with VoIP networks, they'll have a massive retraining job on their hands.


This is exactly what you said:

<>Just look on this board and see how untrainable an engineer is; now put yourself in the RBOC position and imagine training everyone on a whole new IP-based network without completely losing reliability and maybe even service at all.

The meaning of these words are clear. The carriers ha no one trained in IP since they will have to train everyone and the carriers have no IP networks since the IP-based network will be "a whole new" network.

Lewis Carrol's Red Queen comes to mind. Words to her mean what she means them to mean and no more.
dljvjbsl 12/5/2012 | 2:03:05 AM
re: Mobile Skype: Quality Issues?
You need to read more carefully. I did not say that the carriers have no one trained to run IP networks. What I said is that, if they shitcan the PCM networks and replace them with VoIP networks, they'll have a massive retraining job on their hands.


This just does not make sense. However to take it serously for a moment, what did the carriers do, when they replaced their analog networks with TDM digital (in your terms PCM)?
technonerd 12/5/2012 | 2:03:07 AM
re: Mobile Skype: Quality Issues? technonerd you are quite conspiracy theorist.
It is absolutely impossible to be too cynical about a telephone company. You need to remember the history of the Regional Bells. Historically, AT&T central did all the engineering and cap-x planning. The RBOCs did the politics.

The RBOCs have had to develop engineering talent, but their main forte is still politics. If you want to choose to be naive about them, go right ahead, but the reality is that they are as devious as devious gets.


You also do an amazing job of spinning your speculation (some of which is actually accurate) into complete truth.
Just giving my point of view, technoboy. Got a problem with that?


You are equally good at ignoring the points other people make on this board because it doesn't mesh with your line of thinking.
I don't ignore a whole lot, but I do ignore some stuff in the interest of preventing my head from explording all over my new flat panel screen. The cleaning lady has told me that brains, blood and skull fragments are not part of her job description.


You are obsessed with crapping all over a technology that has been in existince for maybe 8-9 years while the PSTN has been around for a lot longer. I guess objectivity is not part of your vocabulary.
I genuinely disagree on both points. I think IP is great, I really do. I think it's best used for messaging right now, and if it going to be used for full-duplex interactive I think video is a much better bet. I think it's the height of stupidity to reinvent the wheel with voice.

As for objectivity, what could possibly be more objective than wanting the shit to work the way we're told it works? What could possibly be more objective than picked up the phone, hearing "Hmmmmm," dialing some numbers and talking to a fellow nutcase 5,000 miles away?


Us propellerheads who have been in the business for a long time are quite aware of the technical issues surrounding VOIP.
Actually, it's "we propellerheads," but I digress. If you're so aware of the technical issues, then why all the goddamn evangelism for a science project still in progress? My "obsession" is with bursting the hype bubbles that have allowed so many California shysters to team up with so many New York shysters to steal so much money.


We have alot of work to do make it work on the scale that is required and the level of security and QOS necessary.
Well, congrats. At least you admit it. Now.


Your premise that the Service providers are doing this to escape regulation is really very shortsighted but I know you dont like to worry about the details.
We all have our islands of competence. I dare say that I'd be happy to put my detailed knowledge of telecom regulation up against that of anyone on this site, or on Light Reading's staff, for that matter. I will forget more than you or they will ever know when it comes to how telecom is regulated.


Especially when they dont support your statements. Continue to rant and rave Mr technonerd. It must be hard not to have any hobbies when your retired.
Oh, but I do have hobbies. I am a shit-missilier. What should I do? Play with model trains?


One of your statements regarding how carriers dont have anyone trained to run an IP network is especially funny. Since each of these carriers have rather large IP networks already and have been offering services on these networks for quite some time I am curious to know who runs these networks.
You need to read more carefully. I did not say that the carriers have no one trained to run IP networks. What I said is that, if they shitcan the PCM networks and replace them with VoIP networks, they'll have a massive retraining job on their hands.

And that's the truth, technoboy. And if they have to do this, it won't be pretty. Now, as much as I distrust phone companies, let me tell you a dirty little secret. All their crap is worth it because they provide cast-iron service. They know that uptime is the only thing that stands between them and regulatory oblivion.

So if you really think they're going to take this cheap (well, at least largely paid-for) network that actually works and replace it with a propellerhead science project, then I want you to tell me where you've been picking your mushrooms. Come to think of it, maybe I could use another hobby.
dljvjbsl 12/5/2012 | 2:03:08 AM
re: Mobile Skype: Quality Issues?
It occurs to me that what really pisses off the IP crowd is that the PSTN works so well. Has anyone noticed how they actually attack its reliability, as if there's something wrong with having a phone that you know will go "hmmmmmmmmmmmmm" when you pick it up at 3 a.m. to call an ambulance because your wife has stopped breathing.

Who is attacking the reliability of the PSTN?

As I stated previously the digital network that you are so enamoured of had to have the reliabilty specifications that were created in the days of electromechanical and electronic switching, modifed so that they could be met with an economical system. Who actually suggesting that network reliability is a bad thing and should be reduced for IP services?

Or is this just another 'shit missile?'
technoboy 12/5/2012 | 2:03:08 AM
re: Mobile Skype: Quality Issues? It occurs to me that what really pisses off the IP crowd is that the PSTN works so well. Has anyone noticed how they actually attack its reliability, as if there's something wrong with having a phone that you know will go "hmmmmmmmmmmmmm" when you pick it up at 3 a.m. to call an ambulance because your wife has stopped breathing.

Who is attacking the reliability of the PSTN?

As I stated previously the digital network that you are so enamoured of had to have the reliabilty specifications that were created in the days of electromechanical and electronic switching, modifed so that they could be met with an economical system. Who actually suggesting that network reliability is a bad thing and should be reduced for IP services?

That is exactly what it is!!!
technoboy 12/5/2012 | 2:03:09 AM
re: Mobile Skype: Quality Issues? technonerd you are quite conspiracy theorist.

You also do an amazing job of spinning your speculation ( some of which is actually accurate) into complete truth. You are equally good at ignoring the points other people make on this board because it doesn't mesh with your line of thinking. You are obsessed with crapping all over a technology that has been in existince for maybe 8-9 years while the PSTN has been around for a lot longer. I guess objectivity is not part of your vocabulary. Us propellerheads who have been in the business for a long time are quite aware of the technical issues surrounding VOIP. We have alot of work to do make it work on the scale that is required and the level of security and QOS necessary. Your premise that the Service providers are doing this to escape regulation is really very shortsighted but I know you dont like to worry about the details. Especially when they dont support your statements. Continue to rant and rave Mr technonerd. It must be hard not to have any hobbies when your retired.

One of your statements regarding how carriers dont have anyone trained to run an IP network is especially funny. Since each of these carriers have rather large IP networks already and have been offering services on these networks for quite some time I am curious to know who runs these networks.
dljvjbsl 12/5/2012 | 2:03:09 AM
re: Mobile Skype: Quality Issues?
I agree fully that the network access should be IP --- single access pipe for all services, ease of extension for network and CPE-based services, but I have never been fully comfortable with the implicit assumption that because the access/edge is IP-based then the core network must also be IP.


I think this is in the cards in the near- to medium-term. It will be caused by a mixture of FCC policy, Wall Street corruption and the stupidity and/or corruption of the media.


I remember an old Dick Van Dyke show program in which Dick was running for city council. Laura asks Dick about the views of his chief opponent. Dick replies (paraphrasing) "He knows what he is talking about. He isn't as free to be as creative as those of us who don't."

He doesn't even understand the implications of the text that he quotes. Everyone is stupid and/or corrupt but him.
technoboy 12/5/2012 | 2:03:09 AM
re: Mobile Skype: Quality Issues? technonerd: Well, look. At the moment, I can pick up my phone, punch in some numbers and call anywhere in the world. No one will hack or hijack my phone, and I don't have to go through an elaborate setup process. And when the call goes through, I'll be able to hear the other guy over in Timbuktu.

Amazing trick. I just picked up my IP phone and called Europe and it worked as well. I didnt dial any special digits or scan my phones for viruses or cast some magic spell. I just picked up the phone and dialed. No black magic technonerd. Feel free to launch your shit missile now!!!!
dljvjbsl 12/5/2012 | 2:03:10 AM
re: Mobile Skype: Quality Issues?
one Fast Ethernet interface on a cheap server can generate thousands of call VoIP call requests per second.


Which will be ignored by the proxy to which they are sent
technonerd 12/5/2012 | 2:03:12 AM
re: Mobile Skype: Quality Issues? I agree fully that the network access should be IP --- single access pipe for all services, ease of extension for network and CPE-based services, but I have never been fully comfortable with the implicit assumption that because the access/edge is IP-based then the core network must also be IP.

I think this is in the cards in the near- to medium-term. It will be caused by a mixture of FCC policy, Wall Street corruption and the stupidity and/or corruption of the media.

The Regional Bells desperately want to kill the Telecom Act. Because there are few competitors left -- the CLECs having been most California/Wall Street frauds, now out of business -- the Regional Bells have forced all of the vendors to become their lobbyists in Washington.

They have convinced the FCC to exempt from the Telecom Act any circuit capable of broadband transmission -- defined by the FCC as anything 200 Kb/s or above. Voice on that circuit, if part of an IP service, will be unregulated. Voila! Voice over IP!!

Now here's the rub. The Regional Bells have a butt-load of Class 5s and Class 4s and STPs and all the other bells and whistles of the PSTN. Most of that crap is paid for. And guess what? It works. And their people are trained on it. Just look on this board and see how untrainable an engineer is; now put yourself in the RBOC position and imagine training everyone on a whole new IP-based network without completely losing reliability and maybe even service at all.

The answer is easy. Channelized Voice over DSL. You establish a PCM tunnel through the lower frequencies, and route those packets over to the PSTN. CVoDSL is already part of all new ADSL chips. It will be very easy to implement this.

Question: What happens when people find out about it?

Answer: Who says they have to? It's a technical issue. Have Vonage do an IPO and have an RBOC buy them; this will buy off the VCs and the Street. The trade journals will say what they are told to say, and the general media will never catch on. We have a press that never met a p.r. handout it wouldn't reprint, and if there's one thing Cisco and the RBOCs have it is big p.r. staffs.

Conclusion: We will probably soon wind up with something that is called "VoIP" or maybe a new 'n snazzier marketing term, but will in reality be the same old telephone call going through the same old switch. The only difference: No regulation of the phone company. Oh joy. Can't you just wait for those $100 a month phone bills?
technonerd 12/5/2012 | 2:03:12 AM
re: Mobile Skype: Quality Issues? Note: fgoldstein, my previous shit-missile wasn't aimed at you. Your remarks inspired me to aim a new one at others.
technonerd 12/5/2012 | 2:03:13 AM
re: Mobile Skype: Quality Issues? The call-rate issue is a serious one. It takes finite time for a call to be initiated using the PSTN; even ISDN PRI has a limit to how much damage can be done over one 64 kbps D channel. The Interent is fast; one Fast Ethernet interface on a cheap server can generate thousands of call VoIP call requests per second. That is just one example of a widely-ignored vulnerability.

IP networking, even for store & forward messaging, is still very much a work in progress. It has been barely used at all for full-duplex, real-time traffic. Yet the same people who gave us the Permanent Boom and the Internet Miracle now come and tell us that the PCM system is dead.

Anyone who believes it should come to me for shares in the Brooklyn Bridge. O.K., there aren't any earnings on the Brooklyn Bridge but who cares? Profits don't count, right? We're talking about a revolution here, and silly technonerd wants to know whether the shit works and when it will make money? What a clueless Philistine he is!

Put a million or so people on residential VoIP, and sure as day follows night it will be a SNAFU a week. Every last one of which will be blamed on the users. All of them. The high-tech people are no different than George W. Bush in this regard. The one thing they will never, ever admit is a "mistake." It's always someone else's fault, or if that won't fly then it's "in the past" and we have to "think of the future."

Well, look. At the moment, I can pick up my phone, punch in some numbers and call anywhere in the world. No one will hack or hijack my phone, and I don't have to go through an elaborate setup process. And when the call goes through, I'll be able to hear the other guy over in Timbuktu.

It occurs to me that what really pisses off the IP crowd is that the PSTN works so well. Has anyone noticed how they actually attack its reliability, as if there's something wrong with having a phone that you know will go "hmmmmmmmmmmmmm" when you pick it up at 3 a.m. to call an ambulance because your wife has stopped breathing.

If I were the California IP neo-fraud crowd, I'd back off of that little tactic as fast as I could. They might not like "five-nines," but I can tell you that normal people really appreciate having telephones that always work. And it really helps that they're easy to use. I hate to break it to the data geeks, but most people are very UN-fascinated by this stuff. They want to pick up and make the goddamned call and have it go through, first time every time.

Propellerheads, may I humbly suggest that you remember this? I know I am harsh now. It hurts me as much as it hurts you for me to be so mean. But trust me, you will thank me when you grow up.
fgoldstein 12/5/2012 | 2:03:14 AM
re: Mobile Skype: Quality Issues? This is a great thread! I'm sorry that I haven't been following it more closely all along. Technonerd's postings have made my day. :-)

It certainly is possible to work VoIP into the network, here and there, almost transparently to the end user. PacketCable is an example. But that's a long way from saying that VoIP will completely replace TDM, or should. I still think TDM is a lot more cost-effective, especially when implemented with modern hardware, for most applications.

The call-rate issue is a serious one. It takes finite time for a call to be initiated using the PSTN; even ISDN PRI has a limit to how much damage can be done over one 64 kbps D channel. The Interent is fast; one Fast Ethernet interface on a cheap server can generate thousands of call VoIP call requests per second. That is just one example of a widely-ignored vulnerability.
dljvjbsl 12/5/2012 | 2:03:19 AM
re: Mobile Skype: Quality Issues?
(b) any failure in transit trunking should be restored in 50ms (SONET/SDH) protection switching times. Good luck with today's IP/MPLS routers. See (a) above.


I think that the issues stated above are very important and deserve serious consideration. However, I was involved in the design of the first digital switches. The same issues about reliability and availability calculations were faced then.

What had seem to happen was that the reliability specs were based not on what was required for service but upon a subliminal understanding of how the then current network worked. When these reliability specs were applied to digital systems, the results were appalling. It did not appear that digital systems could ever meet the reliability requirements for any reasonable degree of cost.

What people realized then was that most if not all of the 'outages' that were being placed on digital systems were in no way customer affecting. The digital systems were being downgraded for 'outages' which would be undetectable for even subscribers who were on-line at the time of 'failure'. The reliability specs were changed (not downgraded) to reflect exactly what they were intended to achieve and not to reflect the operation of electro-mechanical networks.

The 50ms spec indicated above is based on the performance of the human ear. Outages briefer than 50ms are severly attentuated by the human hearing system and so will not be audible.

This does open the question of the relavence of this spec in terms of a network in which msot of the traffic is not voice. 50ms may be far too long for some services such s video and immaterial to other services such as IM. To my mind, it is an open question as to whether or not this spec is valid in the case of a multi-service network.

In any event, relaibility specifications should be designed to serve the needs of the network and its users. They are not absolute statements but must be reviewed in light of the technological and service environments which they serve.
lucifer 12/5/2012 | 2:03:20 AM
re: Mobile Skype: Quality Issues? Stephen - an excellent summary. Let me address some of the points in your posting.

Security -- many carriers are planning to use IP VPNs to separate user traffic to minimise the potential for a malicious user -Što interfere with other users' VoIP and signalling traffic. Communications between Softswitch controllers are encrypted and authenticated and will generally themselves use a separate control plane VPN. There are many policing and "pacing" methods used by SIP Softswitches to limit the potential for Denial of Service attacks.

There is a lot of work going on in the area of "IP CAC" (quite a misnomer since IP has neither the concept of bandwidth or connection) which can be implemented using MPLS mechanisms in the Core IP/MPLS network or by the Softswitch tracking the bandwidth utilisation of calls that it has initiated.

That said, I believe that we don't really know how a large scale VoIP PSTN deployment will function in overload conditions; remember that it has taken over 100 years to get the TDM-based PSTN to the state it now enjoys where users expect to pick up the phone, dial a call, talk and hang-up without congestion issues, poor audio quality or call drops.

IP is possibly the most difficult network protocol that we could have chosen to carry voice services. There are significant challenges with Voice Quality (impact of transit delay, packet delay variation, packet loss) and with end-to-end availability.
The TDM world expects that

(a) all switches should be "Carrier-Grade" -- GR-929 implies that a switch should have a nodal availability of > 99.9999% (yes, 6 nines) for all planned and unplanned outages. The best available routers have downtime during software upgrade and when deployed in pairs to overcome this downtime have protection switching times in the order of seconds for IP and 300 to 500 ms for MPLS (Juniper and Avici offer Virtual APS across two switches which show promise...). The fatal flaw here is the two IP network elements are needed to provide the availability of a single TDM network element

(b) any failure in transit trunking should be restored in 50ms (SONET/SDH) protection switching times. Good luck with today's IP/MPLS routers. See (a) above.

(c) US carriers are required to report service outages or impairments to the FCC. Relatively easy to monitor a TDM connection in that it's either up or down;. how can the network cost-effectively measure the end-to-end performance of millions of VoIP sessions?

In my opinion (and I have worked in the carrier voice area for many years), VoIP is a market-driven solution that has led operators and vendors down the garden path. I agree fully that the network access should be IP --- single access pipe for all services, ease of extension for network and CPE-based services, but I have never been fully comfortable with the implicit assumption that because the access/edge is IP-based then the core network must also be IP.
For example, look at the current PSTN. Access is either analog (loop supervisory signalling, DTMF control signalling) or via ISDN PRI (TDM voice with Q.2931 signalling) but the Core PSTN uses TDM for Voice transport and switching and ISUP for signalling.

If we applied that same logic that we use for VoIP, we should be using analog R2 and R5 signalling and analog voice (as we used to) or use ISDN signalling throughout the whole PSTN (works OK for smallish private networks but does not scale to PSTN sizes)

Of course we cannot ignore the market and I am now happily engaged in a number of projects to get VoIP to as close as possible to the performance levels expected by PSTN users. Lots of work for IP/MPLS router vendors, Softswitch vendors, Gateway vendors, ...

Lucifer
technoboy 12/5/2012 | 2:03:22 AM
re: Mobile Skype: Quality Issues? technoboy:This reply is very telling because it shows your lack of understanding. Network operators dont provide QOS and security. The network infrastructure does that.

technonerd: And YOU are the one who called this discussion "insane?"

Technoboy replies: Why?
dljvjbsl 12/5/2012 | 2:03:23 AM
re: Mobile Skype: Quality Issues?
Presuming that these mechanisms will be located on the carrier side of the demarc, would this not tend to be yet another nail in the coffin of Isenberg's "Stupid Network" thesis? Sounds to me as if the IP carrier is going to have to be at least as deterministic at the PCM carrier; this will require considerable intelligence in the network.


I know that many people have tried to explain this to you; however I will give it one more try.

The idea that there would be intelligence at the carrier periphery (your "carrier side of the demarc")is right out of Isenberg's ideas. Isenberg was contrasting this type of idea with the current AIN in which intelligence is centralized in SCPs and in which the network is designed with assumptions as to the type of traffic that it will carry.

A stupid network will allow very intelligent and adaptive applications by getting out of their way.

Nothing that you have said gives me any indication that you understand what Isenberg is talking about
technonerd 12/5/2012 | 2:03:27 AM
re: Mobile Skype: Quality Issues? This reply is very telling because it shows your lack of understanding. Network operators dont provide QOS and security. The network infrastructure does that.

And YOU are the one who called this discussion "insane?"
technoboy 12/5/2012 | 2:03:28 AM
re: Mobile Skype: Quality Issues? Production means in a real network with real equipment. Not reading some nonsense off of a web site.

technonerd wrote:The banking example is a case where the bank's server is protected, and the transmission is encrypted. The network operator isn't involved. In VoIP, the network operator will need to provide security and QoS, along with a bunch of other functions.

This reply is very telling because it shows your lack of understanding. Network operators dont provide QOS and security. The network infrastructure does that. Just as someone has to configure network infrastructure someone also has to configure the server to provide secure banking.
technoboy 12/5/2012 | 2:03:29 AM
re: Mobile Skype: Quality Issues? For the record I think most of these posts are insane. Unfortunately most people have mis-interpreted the architecture and have made fundamentally poor observations on how these networks are being built. When someone attempts to inject some logic into the conversation we get zombie machines and shit missiles as a response. Lets just all stop what we are doing and stick with the good old PSTN (Which by the way is subject to hundreds of millions of dollars in toll fraud annually by many unscrupulous individuals) and yet it manages to continue to provide dial tone. Yet you are all absolutely sure that the sky will fall if we do this over IP. Is everyone really this dense (except you technonerd as I know you are in this for personal amusement only). Do not call lists cant exist in an IP communications networks. Everyones IP phone will be highjacked. Has anyone on this email thread even used the techonology in a production environment. You guys really crack me up. Alchemy I would bet you your entire salary you could not break my SIP proxy with the attack you described. You wouldnt make it past the edge of the network. Sorry no cigar for you. Technonerds constant references to feces leads me to believe he is need of some stool softeners. Stephencooke needs to stop watching "dawn of the dead".
technonerd 12/5/2012 | 2:03:29 AM
re: Mobile Skype: Quality Issues? With telephone banking, one could see a PC with a DTMF sender and voice receiver emptying any number of accounts. With Internet banking one could see the same thing happening, the hacker could transfer money at will.

We do not see these things happening because of the types of security that surround these transactions. These same security measures are being used to protect VoIP.

The banking example is a case where the bank's server is protected, and the transmission is encrypted. The network operator isn't involved. In VoIP, the network operator will need to provide security and QoS, along with a bunch of other functions.


It would be trivial to program a PC to amke an unending stream of abusive hang up calls to a single or group of numbers.
And you can almost instantly trace it back to the caller. Not so with DoS attacks. The analogy would be someone not only programming his own computer to make these calls, but thousands of other computers. And the vulnerability would be located in the IP world; the dialing would merely be a conduit for IP zombies, who grow thick in Northern California.


Unfortunately most people have mis-interpreted the architecture and have made fundamentally poor observations on how these networks are being built. When someone attempts to inject some logic into the conversation we get zombie machines and shit missiles as a response.
Methinks the technoboy doth protesteth too much. The "divergent" responses come as the result of your assumptions being challenged. This is like shooting fish in a barrel with engineers. While I love engineers dearly, they do tend to look at the world through blinders.


Lets just all stop what we are doing and stick with the good old PSTN (Which by the way is subject to hundreds of millions of dollars in toll fraud annually by many unscrupulous individuals) and yet it manages to continue to provide dial tone. Yet you are all absolutely sure that the sky will fall if we do this over IP.
If you want to make telephone calls, I think the PSTN is definitely the way to go. Changing this over to VoIP has nothing whatsoever to do with technical factors; rather, it is all about regulatory arbitrage of one sort or another.

Mixed with a generous helping of geektitude and a hope that some more bullshit IPOs can be thrown out there so sop up all that liquidity that the Federal Reserve has created because it is so panicked about the financial system and the economy.


Has anyone on this email thread even used the techonology in a production environment.
What do you mean by "a production enviornment?" How come you guys can never talk concretely? Talking with data geeks is like picking up mercury. Your words slither everywhere, the only certainty being that when the shit doesn't work it will be the user's fault.


Technonerds constant references to feces leads me to believe he is need of some stool softeners.
Well, you have to admit it's been kind of a shitty week over there in Iraq, too. Speaking of which, I think Bush is on TV. I better, um, run.
dljvjbsl 12/5/2012 | 2:03:30 AM
re: Mobile Skype: Quality Issues?
PCM system is zombie-proof, but the IP system is zombie-friendly. Maybe it was designed by zombies?


As I have said before, DTMF senders on modem are common aspects of PCs now. It would be trivial to program a PC to amke an unending stream of abusive hang up calls to a single or group of numbers.

One person did do this and was convicted of it. He was angry with an evangelist who had taked very alrge donations from his mother and so wanted to abuse the evangelist and stick him with a lrage 800 number bill.

So the 800 number system which is always touted as the example of smart network capability is very highly susceptible to a zombie attack. With 800 number bills, a hacker could bankrupt a small business.
dljvjbsl 12/5/2012 | 2:03:31 AM
re: Mobile Skype: Quality Issues?
How long do you think it will take before some enterprising hackers figure out how to cripple even parts of the PSTN? "Encrypted" and "trusted" proxy mechanisms, when implemented, can only go so far.


To add another observation here, the same could be said about banks. With Internet and telephone banking, there is always teh possibility that enterprising hackers could break into accounts and steal a great deal of money.

With telephone banking, one could see a PC with a DTMF sender and voice receiver emptying any number of accounts. With Internet banking one could see the same thing happening, the hacker could transfer money at will.

We do not see these things happening because of the types of security that surround these transactions. These same security measures are being used to protect VoIP.

SPAM has arisen because of the lack of security in the Email protocol. Identity spoofing is trivial. In VoIP, it is extremely difficult to spoof a FROM address because this is the way that it ahs been designed.
technonerd 12/5/2012 | 2:03:31 AM
re: Mobile Skype: Quality Issues? One thing we definitely know is that the IP telephony system, such as it is, is promoted and defended by zombies. Or, to be more precise, Cali-Zombies. You know, Santa Clata County and all that. The ones who didn't make it during the last Bubble are hanging on by their fingernails. "Oh, please God, give us one more hot IPO and I promise I never hire a hooker at Supercomm again!"
technonerd 12/5/2012 | 2:03:32 AM
re: Mobile Skype: Quality Issues? "Do Not Call" is something that can be done quite easily on an IP system. Cryptographically secure authentication mechanisms have been designed for SIP. These will be used to vet calls before they can be initiated and to verify calls before they can be terminated.

Presuming that these mechanisms will be located on the carrier side of the demarc, would this not tend to be yet another nail in the coffin of Isenberg's "Stupid Network" thesis? Sounds to me as if the IP carrier is going to have to be at least as deterministic at the PCM carrier; this will require considerable intelligence in the network.


So if a IP telephone is taken over by a zombie and begins to make abusive calls, the calls can be intercepted and denied at both the incoming and outgoing proxies.

But will they be intercepted? And why should there even be a possibility that an IP telephone will be taken over by any zombie other than the subscriber who is using it? The PCM system is zombie-proof, but the IP system is zombie-friendly. Maybe it was designed by zombies?
dljvjbsl 12/5/2012 | 2:03:32 AM
re: Mobile Skype: Quality Issues?
Good God, are you still harping because I failed to write a dissertation on how PBXs work? What is it that you want to know? By the way, I'll freely acknowledge that I'm not a propellerhead and that I know more about some telecom subjects than others. A stunning admission! That said, I don't think my ignorance of PBXs is nearly as complete as you have implied.


You have called enterprise communication sytems/PBXs glorified intercoms. PBXs and enterprise communication systems in general are as complex to build and have a greater feature set than anything in the PSTN. Do you actually think that the communication system that a company relies on for its day-to-day business to connect to its customers and suppliers, can in any way be compared to a glorified intercom?

dljvjbsl 12/5/2012 | 2:03:33 AM
re: Mobile Skype: Quality Issues?
The impact of "Do Not Call" does not exist on an IP-based network. ANY kind of automated calling, be it Audio SPAM, "heavy breathing", Presidential messages, etc. now has the potential to call every household in New York (or any other city in the world) for hours or days on end. If you do not know the originating phone number of the call, how do you stop it? If it is the output of an Internet-based worm, resident on thousands or millions of unwitting home computers (zombies) CAN you stop it?


"Do Not Call" is something that can be done quite easily on an IP system. Cryptographically secure authentication mechanisms have been designed for SIP. These will be used to vet calls before they can be initiated and to verify calls before they can be terminated.

So if a IP telephone is taken over by a zombie and begins to make abusive calls, the calls can be intercepted and denied at both the incoming and outgoing proxies.

technonerd 12/5/2012 | 2:03:34 AM
re: Mobile Skype: Quality Issues? Technonerd: VoIP is a propellerhead science project that needs to be hit with a shit missile (Sorry, I paraphrased here, he said many good things in his posts).

No need to apologize. You paraphrased quite accurately.


If it is the output of an Internet-based worm, resident on thousands or millions of unwitting home computers (zombies) CAN you stop it? If it doesn't overload a gateway or a class 5 in terms of processing power but uses up all available ports how can you guarantee that people on that class 5 will still get dialtone? If they cannot get dialtone how can they call 911? What if the perpetrators of this kind of thing focus on groups of 100 phone numbers in any one city? This will not overload a class 5 but will surely kill ANY meaningful phone service that these people have.

Inevitable answers from propellerheads when it happens:

- The stupid newbie waste-of-skin users are at fault.

- There will inevitably be potholes on the informational superhighway.

- Those people would have died anyway, eventually.


I have now heard several people say that new, value-added services are going to drive VoIP deployment. Will these services or operating efficiencies protect us from what I have described or will they enable it? Are those services really worth the risk?

Outstanding post as usual Stephen. My question is even more basic. What [i]ARE[/i] the new services? It would be nice to have some sort of idea before we stick everyone's gonads on the chopping block. Just a thought.
stephenpcooke 12/5/2012 | 2:03:34 AM
re: Mobile Skype: Quality Issues? This has been an amazing discussion of technology "can" and "can'ts", switch and gateway capabilities, and what the vision is for newer and "better" network architectures when/if they are implemented. A few exerpts:

Me: "As long as 'virtual', non-human-initiated phone calls can be generated by (put your favorite Internet-based nasty here, making sure that you include viruses, adware, spyware, worms, etc.)-infected computers, the POTS network is vulnerable."

dljvjbsl: "The issue on that has been decided. The curent toll network is being replaced by an IP network based on MPLS. This debate seems to be over until the next technological dislocation takes place."

alchemy: "With a bare-naked SIP Proxy (typically a Sun Netra), it's quite easy to mount a denial of service attack that will kill it. All I have to do is nail it with 5000 REGISTER messages per second from a bunch of zombie machines and I've created a doorstop."

PO: "If an individual line goes on/off hook enough over a particular time interval, the Class 5 should start to ignore that line and run diagnostics to see what's wrong. But if a large number of lines originate calls simultaneously then the CPUs in the Class 5 will start to deny originations. This is because these things were designed when memory & horsepower were expensive, so there simply isn't enough capacity to handle everything happening at once. Finally, of course, if the trunks are all occupied then again new originations will fail."

Aswath: "Also keep in mind that in PSTN switches are king. They can cut off any call at any time. They have rules to handle focused overloads and the like. So the mantra is: There are no additional operational concerns introduced because of interworking with VoIP."

Technonerd: VoIP is a propellerhead science project that needs to be hit with a shit missile (Sorry, I paraphrased here, he said many good things in his posts).

As I have also said, I believe that VoIP has opened Pandora's box in ways that we cannot yet imagine. Never before has the PSTN been accessible to viruses, worms, zombies, etc. It is now. How long do you think it will take before some enterprising hackers figure out how to cripple even parts of the PSTN? "Encrypted" and "trusted" proxy mechanisms, when implemented, can only go so far. The impact of "Do Not Call" does not exist on an IP-based network. ANY kind of automated calling, be it Audio SPAM, "heavy breathing", Presidential messages, etc. now has the potential to call every household in New York (or any other city in the world) for hours or days on end. If you do not know the originating phone number of the call, how do you stop it? If it is the output of an Internet-based worm, resident on thousands or millions of unwitting home computers (zombies) CAN you stop it? If it doesn't overload a gateway or a class 5 in terms of processing power but uses up all available ports how can you guarantee that people on that class 5 will still get dialtone? If they cannot get dialtone how can they call 911? What if the perpetrators of this kind of thing focus on groups of 100 phone numbers in any one city? This will not overload a class 5 but will surely kill ANY meaningful phone service that these people have. Given the potential for millions of home computer users to be infected with something like this what is to stop these hackers from hitting random cities across any country that they choose?

I have now heard several people say that new, value-added services are going to drive VoIP deployment. Will these services or operating efficiencies protect us from what I have described or will they enable it? Are those services really worth the risk?
technonerd 12/5/2012 | 2:03:40 AM
re: Mobile Skype: Quality Issues? I repeat what I have said before you have demonstrated a complete misunderstanding of the functioning and capabilities of PBXs.You really do no know how they work or what they do.

Good God, are you still harping because I failed to write a dissertation on how PBXs work? What is it that you want to know? By the way, I'll freely acknowledge that I'm not a propellerhead and that I know more about some telecom subjects than others. A stunning admission! That said, I don't think my ignorance of PBXs is nearly as complete as you have implied.
dljvjbsl 12/5/2012 | 2:03:43 AM
re: Mobile Skype: Quality Issues? referring to enterprise telephony

If it's a corporate or household firewall, I guess we'll just have to agree to disagree. I likened it to a glorified intercom, but that was an analogy. Maybe the propellerheads have become so thoroughly propaganduized that can't fully appreciate an analogy.


I repeat what I have said before you have demonstrated a complete misunderstanding of the functioning and capabilities of PBXs.You really do no know how they work or what they do. It is hard discussing new models of communication and collaboration with that sort of handicap.
technonerd 12/5/2012 | 2:03:43 AM
re: Mobile Skype: Quality Issues? call forwarding your mobil phone from the US to Europe is about the most ridiculous thing I have heard anyone ever do.

Why?

technonerd 12/5/2012 | 2:03:44 AM
re: Mobile Skype: Quality Issues? I am assuming you put my name in this post so that I would reply.
I did nothing other than click on "Reply."


From your response, it appears as though you would like to put further qualifications on what constitutes a telephone call.
That's an exaggeration.


IP communications behind a firewall is nothing close to a glorified intercom and you know this.
If it's a corporate or household firewall, I guess we'll just have to agree to disagree. I likened it to a glorified intercom, but that was an analogy. Maybe the propellerheads have become so thoroughly propaganduized that can't fully appreciate an analogy.


You are making this statement because the enterprise is adopting the technology and that infringes on your statements that no one uses it which is complete nonsense.
Way back when in this discussion, I stated that there VoIP was a niche service, and that one of its niches was some large enterprises.
dljvjbsl 12/5/2012 | 2:03:45 AM
re: Mobile Skype: Quality Issues?
According to Broadband Reports bulletin board (http://www.broadbandreports.co..., AT&T CallVantage uses MGCP. If it is so, then intelligence is not at the end-points; the architecture is equivalent to ISDN BRI with stimulus signaling. All features and services must be developed at the Call Agent. But unfortunately I can not officially verify because AT&T is quiet about its architecture.


This is a standard VoIP architecture and is completely compatible with SIP.

The Media Gateway resides at the customer site and is connected by MGCP to the Media Gateway Controller (MGC) at the carrier end.

Features could be defined on the SIP proxy whcih presumably resides within the MGC. Features are defined locally at the SIP proxy and negotiated end to end with the proxy that serves the distant party.

This is a pure SIP VoIP architecure


There is a good reason why features are defiend this way and not at the customer telephone. SIP and VoIP ar desihgned to support user mobiloity. Users are not confined to a single device and may have multiple owned devices and can register on other devices in places wher they are visiting. The SIP Register command allows for this. Thus no suer device and claim premiince in serving the user. The user may be at any device.

The place where his features are defined and operated must always be powered (which is not the case ofr all end user deivices). The proxy will reside at the carrier site for publlic services and in a similar private site for enterprise services.

Users will have only one number which will direct callers to the proxy. The calledproxy will decide in concert with the callers proxy how the call should be handled.

Note that this is completely different than the system defined by the AIN (Advanced Intelligent Network) which Isenberg was talking about. In this system all features are defined and operated centrally. This system has its advantages but customization, evolvability, convergence, colaboration and personalization are not among tem.
technoboy 12/5/2012 | 2:03:45 AM
re: Mobile Skype: Quality Issues? Aswath: Here is a link to their overall architecture and strategy. It took about 30 seconds to find it on their website. You are entitled to interpret the architecture any way you like but it is your interpretation and not necessarily accurate. Again I ask someone to define the word " network" what does it include?

http://www.att.com/attlabs/dev...

In the case of Vonage I am not seeing your point. The managed device is to the ATA. You are plugging your phone into that device that they provide. It is essentially a piece of CPE or in this case the endpoint device. Since the service is provided over a broadband connection the bandwidth is already there. There are other customers that are trialing SIP phones at their locations. Tell me what service providers dont let you run the services at the endpoint.
aswath 12/5/2012 | 2:03:47 AM
re: Mobile Skype: Quality Issues? look at AT&T overall architecture and then ask the same question.

As I mentioned in the previous post, I can not look at the overall architecture. From what I gather, it uses MCGP; if that is the case, then I feel it does not adhere to "stupid network" philosophy. If you know otherwise, can you please elaborate?

Finally SIP defines features that run on the endpoint and in the system so yes if they provide the CPE then they provide the service.

SIP may allow features to be run at both the places; but many service providers do not allow the subscriber to run their own applications. For example can we run our own application on top of Vonage's ATA? Also, according to Vonage itself, 3-way bridging is taking place in the ATA; still it claims it offers 3-way calling. Shouldn't this be challenged? In PSTN, with a single access line I can have a 3-way call because the service provider offers; in Vonage's case, I better have sufficient bandwidth to carry two independent calls; Vonage does not do anything extra.

I agree that in IP networks, intelligence can be moved to the edges - not because Isenberg says so, but because Padlipsky and Clark said so. But many of the VoIP providers are blocking it in many subtle ways without being challenged.
aswath 12/5/2012 | 2:03:47 AM
re: Mobile Skype: Quality Issues? I have only see the press pieces on this service. From them, I did not see naything that that makes this service different than any other VoIP service. What am I missing?

According to Broadband Reports bulletin board (http://www.broadbandreports.co..., AT&T CallVantage uses MGCP. If it is so, then intelligence is not at the end-points; the architecture is equivalent to ISDN BRI with stimulus signaling. All features and services must be developed at the Call Agent. But unfortunately I can not officially verify because AT&T is quiet about its architecture.
technoboy 12/5/2012 | 2:03:50 AM
re: Mobile Skype: Quality Issues? aswath:Have you looked at AT&T CallVantage? Do you think it adhers to "stupid network" philosophy? I am also interested in your views on some of the SIP-based service providers that claim to offer features like Call waiting and 3-way calling when it is really done by the CPE.

Well aswath look at AT&T overall architecture and then ask the same question. Also define network in a meaningful way. Finally SIP defines features that run on the endpoint and in the system so yes if they provide the CPE then they provide the service.
technoboy 12/5/2012 | 2:03:51 AM
re: Mobile Skype: Quality Issues? technonerd wrote:Service providers are moving to a single architecture and that is IP.
That is a completely meaningless statement. Do you mean "all service providers?" What is your definition of "moving to?"

They are building these networks now (AT&T, MCI, Sprint, Verizon, Qwest, BT, Deutche Telecom, Bellsouth, SBC, and many others) If VOIP( now to be known as IP COMMUNICATIONS) is a geek toy than call forwarding your mobil phone from the US to Europe is about the most ridiculous thing I have heard anyone ever do.
technoboy 12/5/2012 | 2:03:51 AM
re: Mobile Skype: Quality Issues? alchemy wrote:Why would the signaling protocol make one lick of difference as to the vulnerability of a media gateway to denial of service attacks. Doesn't matter if it's SIP, MGCP, MEGACO, or smoke signals. If you can reach it from your IP phone sitting on your residential broadband connection and you're not going through VPN or a session controller/border proxy, the media gateway can be attacked with a distributed denial of service attack. It wouldn't be too difficult to blow away Vonage service, for example.

First of all I claimed nothing. I asked a question about where this has occurred and since you did not respond to the actual question I am left to assume it has not occurred.

Alchemy do me a favor and read all of my posts. It makes a big difference. First of all Vonage does not provide a service based on MGCP. Second of all the MGCP architecture is very different than one based on SIP which incorporates things like session controllers and proxies. Third, the platforms that service providers are offering or not based on customer facing media gateways which are then connected to a class 5. I will say it for the last time. DDOS, Viruses, trojan horses, worms and whatever else you can think of will not change the fact the service providers are still going to build these networks and provide their services over IP. You need to be more specific with your examples. I mean anyone can cut and paste crap off of an internet site paste it into one of these posts.
technoboy 12/5/2012 | 2:03:52 AM
re: Mobile Skype: Quality Issues? technonerd wrote:And if we're talking about working great behind the corporate firewall, what we're actually talking about isn't really telephone service but a glorified intercom. It's only a telephone if you can make and receive calls to and from anyone with a phone. That's what the PCM network does. The IP network is a science experiment yet incapable of doing it.

Class 4 is only one piece of the puzzle. To cite Internet offload as an example of IP's success, well, all I can do is ask the following question: Is it possible to puke and laugh at the same time?

I am assuming you put my name in this post so that I would reply. Since I did not write the original post I am not sure what the poster was attempting to convey. From your response, it appears as though you would like to put further qualifications on what constitutes a telephone call. I have stated many times although no ones wants to realize it that VOIP or IP communications over the public internet is a small piece of the puzzle. Service providers are building their own IP networks and providing these IP communications services over that network. IP communications behind a firewall is nothing close to a glorified intercom and you know this. You are making this statement because the enterprise is adopting the technology and that infringes on your statements that no one uses it which is complete nonsense. Obviously when you call off net you are going to pass through gateway into the PSTN. I have stated this many times. There is a necessary co-existence between POTS and IP communications.

Internet offload did work very well and freed up alot of space on thos precious carrier switches. Thats not really what I have been referring to as that architecture talks to MGCP and the future is really SIP.
dljvjbsl 12/5/2012 | 2:03:52 AM
re: Mobile Skype: Quality Issues?
Have you looked at AT&T CallVantage? Do you think it adhers to "stupid network" philosophy


I have only see the press pieces on this service. From them, I did not see naything that that makes this service different than any other VoIP service. What am I missing?
technonerd 12/5/2012 | 2:03:55 AM
re: Mobile Skype: Quality Issues? The success stories for VoIP are for applications where the service sits on a managed IP network isolated from this sort of attack. Class 4 and internet offload work great. Enterprise IP PBX behind the corporate firewall works fine though you can certainly attack the VPN service and deny telephony service to anybody outside the firewall.

And if we're talking about working great behind the corporate firewall, what we're actually talking about isn't really telephone service but a glorified intercom. It's only a telephone if you can make and receive calls to and from anyone with a phone. That's what the PCM network does. The IP network is a science experiment yet incapable of doing it.

Class 4 is only one piece of the puzzle. To cite Internet offload as an example of IP's success, well, all I can do is ask the following question: Is it possible to puke and laugh at the same time?
technonerd 12/5/2012 | 2:03:55 AM
re: Mobile Skype: Quality Issues? I would really prefer a more open and intellectual debate so that we can really get into the issues and obtain a better understanding of what is going with networks and applications and all that other stuff.

I have been "open" and "intellectual" throughout, my colorful language notwithstanding. When I encournter meaningless hype, I launch a shit-missile at it. I will continue to do so.
technonerd 12/5/2012 | 2:03:55 AM
re: Mobile Skype: Quality Issues? I mean lets be realistic. IP communications is not a toy.
I have never said it was.


Many companies have adopted various pieces of the technology. Everything from email to ERP to video conferencing to running their voice over their IP infrastructure.
Email is useful. ERP, which predated the Internet and IP, is useful. Videoconferencing is useful when done right. VoIP is a geek toy.


Service providers are moving to a single architecture and that is IP.
That is a completely meaningless statement. Do you mean "all service providers?" What is your definition of "moving to?"
aswath 12/5/2012 | 2:03:56 AM
re: Mobile Skype: Quality Issues? The Internet introduces a completely different and more effective model.

Have you looked at AT&T CallVantage? Do you think it adhers to "stupid network" philosophy? I am also interested in your views on some of the SIP-based service providers that claim to offer features like Call waiting and 3-way calling when it is really done by the CPE.
alchemy 12/5/2012 | 2:03:57 AM
re: Mobile Skype: Quality Issues? technoboy claims:
Tell me where this has occurred and also you are referring to a media gateway which using MGCP or perhaps megaco.

Why would the signaling protocol make one lick of difference as to the vulnerability of a media gateway to denial of service attacks. Doesn't matter if it's SIP, MGCP, MEGACO, or smoke signals. If you can reach it from your IP phone sitting on your residential broadband connection and you're not going through VPN or a session controller/border proxy, the media gateway can be attacked with a distributed denial of service attack. It wouldn't be too difficult to blow away Vonage service, for example.

The success stories for VoIP are for applications where the service sits on a managed IP network isolated from this sort of attack. Class 4 and internet offload work great. Enterprise IP PBX behind the corporate firewall works fine though you can certainly attack the VPN service and deny telephony service to anybody outside the firewall.
technoboy 12/5/2012 | 2:03:58 AM
re: Mobile Skype: Quality Issues? I was referring to the discussion of his book as it has been related to the internet or the IP network or whatever else we want to call it. Many people have put forth there own interpretation of his book and I am no exception. I am stating now that many posts on this site are way too ambiguous and when challenged we like to reply to the things we are comfortable with. I am probably guilty of this as well. I would really prefer a more open and intellectual debate so that we can really get into the issues and obtain a better understanding of what is going with networks and applications and all that other stuff. Just my thoughts before I retire for the evening.
technoboy 12/5/2012 | 2:03:58 AM
re: Mobile Skype: Quality Issues? technonerd wrote:The new term is fuzzier than the one it replaces. VoIP is a subset of IP communications. I suspect that people who use it do so because they know the VoIP is a geek toy and want to be taken more seriously than they otherwise would be.

The new term reflects the reality that all types of media are being adapted to IP. Whether or not you like it is not relevant.

Yes " Shit Missile is quite Humurous"

I have posted the statement on Isenberg to several people on this board. It is just a dumb argument to be having. I have also stated that we should all be trying to make our points just a tad bit clearer and responding to specific questions and not just making statements we like to make. All this other stuff is to coin a phrase " a big nothingburger".

I mean lets be realistic. IP communications is not a toy. Many companies have adopted various pieces of the technology. Everything from email to ERP to video conferencing to running their voice over their IP infrastructure. Service providers are moving to a single architecture and that is IP. It will take a long time and there will be problems along the way but all this talk about DDOS attacks and trojan horses and whatever else are side bar conversations at best. If these issues were that detrimental we would not be moving in this direction. These are serious issues but they are being addressed in the natural maturation of the technology.
dljvjbsl 12/5/2012 | 2:04:00 AM
re: Mobile Skype: Quality Issues?
This whole issue of isenberg's book on the rise of the stupid network is not worth discussing because it leaves too much open to interpretation. What is a smart network? what is a dumb network? What layer defines intelligence? What makes up the network? Is it transport? Application? Network management? Hard to decipher and equally unimportant!


Isenberg was quite specific about this. he was reacting to the attempt to build the 'Advanced Inteligent Network'. The AIN, as we all know, attempts to be intelligent by providing a centralized processing capacity. Centralization was supposed to ease the development of new functionality by limiting the amount of change required to introduce a new feature. The issue on which this design founders is that it cna cope with complexity but it cannot really cope with customization and evolution. As designed it provides a complex set of services to all subscribers. If a group of subscribers wishes to have a different type of service, then they will go on wishing; the AIN has no real answer to this.

The Internet introduces a completely different and more effective model.
technonerd 12/5/2012 | 2:04:01 AM
re: Mobile Skype: Quality Issues? I call it IP Communications and it is a term that is beginning to replace VOIP, Call it semantics if you like
The new term is fuzzier than the one it replaces. VoIP is a subset of IP communications. I suspect that people who use it do so because they know the VoIP is a geek toy and want to be taken more seriously than they otherwise would be.


I think everyone would appreciate that conversation and not another rendition of technonerds greatest one liners.
You have to admit that I can turn a phrase. I like "shit missile," myself.


This whole issue of isenberg's book on the rise of the stupid network is not worth discussing because it leaves too much open to interpretation. What is a smart network? what is a dumb network? What layer defines intelligence? What makes up the network? Is it transport? Application? Network management? Hard to decipher and equally unimportant!
Hey, I've only discussed Isenberg in reaction to others who continue -- amazingly enough -- to promote this guy's so-called ideas.
technonerd 12/5/2012 | 2:04:01 AM
re: Mobile Skype: Quality Issues? I would refer you to the issue of harrassing and obscene calls. This is a DoS attack by any definition.
Yeah, I think I'll simultanouesly call, oh, a couple million woman and breathe real hard. DoS attacks? Come on, get serious or I'll start calling you an idiot.


SMDR (system message detail records) contain all of this information about calls. It also contains much other interesting information such as the telephone credit card that has ben used. Only trustworthy people are allowed to see these records.
And in the IP world, anyone can (and does) attach a tracking device to your browser. You have to buy special devices to unattach those bugs, i.e., anti-spyware, firewalls, etc., and even then we really don't know if it works.

Of course, the "Trust E" organization can be counted on to discipline unscrupulous invasions of privacy in the IP world. Why, just look at all the companies they've publicly censured!


Naturally there are stories, perhaps apocryphal, about people who have stolen this information and used it for fraud
Oh, I'm sure it has happened. But not as a matter of course, that's for damn sure. Although, on the other hand, there is evidence that the phone companies themselves can -- and occasionally do -- sell your calling data. Has anyone noticed that they only started to do it when the IP world made invasion of privacy a routine business practice?
dljvjbsl 12/5/2012 | 2:04:02 AM
re: Mobile Skype: Quality Issues?
No one attaches "cookies" to telephone calls such that the recipient of one call can see all the calls you've made before you called him and all the calls you made afterwards. No one needs to buy virus protection software for their telephone -- unless IP in somehow involved.


SMDR (system message detail records) contain all of this information about calls. It also contains much other interesting information such as the telephone credit card that has ben used.

Only trustworthy people are allowed to see these records. Naturally there are stories, perhaps apocryphal, about people who have stolen this information and used it for fraud,
technoboy 12/5/2012 | 2:04:02 AM
re: Mobile Skype: Quality Issues? technonerd wrote:IP and the networks that carry it are by their very nature insecure relative to PCM. And the technology of real-time IP full-duplex comms remains quite immature, which means that any users will be laboratory rats for the propellerheads in the industry.

This is essentially a false statement that VOIP ( I call it IP Communications and it is a term that is beginning to replace VOIP, Call it semantics if you like) is a toy. It has been proven that many large companies have adopted the technology and it continue to grow far more rapidly than traditional TDM phones systems. Period.

Tehcnonerd:Hell no! One of the conceits of the IP people is that somehow their technologies will eliminate the tendency of human beings to lie, cheat and steal. You'd think that the experience of the Bubble, which is by a huge margin the biggest set of interconnected financial frauds the world has ever seen, would at least make these people a bit more humble.

This is essentially your view and I have not seen anyone else state that IP will solve the worlds problems. You created these statement and make them appear as facts.

Finally, you have ignored the previous posts when I have asked you to defend your statements. Lets have a real discussion on the merits or lack their of VOIP. I think everyone would appreciate that conversation and not another rendition of technonerds greatest one liners.

This whole issue of isenberg's book on the rise of the stupid network is not worth discussing because it leaves too much open to interpretation. What is a smart network? what is a dumb network? What layer defines intelligence? What makes up the network? Is it transport? Application? Network management? Hard to decipher and equally unimportant!!!!

dljvjbsl 12/5/2012 | 2:04:03 AM
re: Mobile Skype: Quality Issues?
Oh, and tell me, is your "telephone firewall" up to date? If you have a VoIP phone, let's hope so. If you're using the PSTN, there's no need now is there?


It is interesting to see that your analysis has the following charactristics:


a) the interesting actvities take place at the periphery of the network

b) attacks on the periphery can be addressed by authentication and authorization techniques


As to the assertion that DoS attacks on POTS phones are not possible, I would refer you to the issue of harrassing and obscene calls. This is a DoS attack by any definition. CLASS services with Call Trap, and other features are used to defend against these types of DoS attacks. This is a description of a telephone firewall in the PSTN network. It uses authentication and authorization techniques.

Note that intelligent IP phones which could detect and bar unwanted callers would seem to be better able to cope with harassing and obscene calls.
technonerd 12/5/2012 | 2:04:06 AM
re: Mobile Skype: Quality Issues? Oh, and tell me, is your "telephone firewall" up to date? If you have a VoIP phone, let's hope so. If you're using the PSTN, there's no need now is there?
technonerd 12/5/2012 | 2:04:07 AM
re: Mobile Skype: Quality Issues? In what way are IP networks less secure than TDM (PCM) networks. The PSTN is secured by isolating signaling from the periphery by filters on the trunks. IP networks are secured in the same way by isolating signalling channels by way of cryptography. I really do not see much of a difference.
For starters, it's impossible to mount a denial of service attack on a PCM system. It's impossible to hijack someone's telephone -- unless it's connected to a computer, in which case the hijacking will take place via an IP network!

SONET transmissions are about a secure as secure gets. True, there are now CALEA servers, but that's a whole lot different than any Tom, Dick or Harry putting a packet sniffer out there to intercept people's e-mail.

No one attaches "cookies" to telephone calls such that the recipient of one call can see all the calls you've made before you called him and all the calls you made afterwards. No one needs to buy virus protection software for their telephone -- unless IP in somehow involved.
dljvjbsl 12/5/2012 | 2:04:08 AM
re: Mobile Skype: Quality Issues?
IP and the networks that carry it are by their very nature insecure relative to PCM.


In what way are IP networks less secure than TDM (PCM) networks. The PSTN is secured by isolating signaling from the periphery by filters on the trunks. IP networks are secured in the same way by isolating signalling channels by way of cryptography. I really do not see much of a difference.
technonerd 12/5/2012 | 2:04:10 AM
re: Mobile Skype: Quality Issues? In my long "summary" post, I called David Isenberg a "shyster." I could imagine him or someone else viewing my use of that particular word as an allegation of personal financial corrpution.

So, let me say that I have no such evidence of that. I used the word figuratively, in the sense that I think Isenberg has propogated an outrageously incorrect thesis. But I did not, and do not, accuse him of corruption.

I should also say that I'm not posting this as the result of having received any communication from Isenberg or anyone else. I am posting it after having given some thought to what I wrote.

Webster's defines shyster thusly: one who is professionally unscrupulous especially in the practice of law or politics.

Webster's notes that the word is probably derived from the German Scheisser. Literally, defecator.

Frankly, seeing as how I am the one who expressed a fondness for launching shit-missiles on this board, I should be calling myself the shyster here? Anyway, if David should happen to see my posting I apologize for any implication of professional misconduct on his part.

I disagree very strongly, but not that strongly, with him.
technonerd 12/5/2012 | 2:04:10 AM
re: Mobile Skype: Quality Issues? I think Light Reading is the only board on the Internet that requires you to use the sideways caret (< >) symbol to do punctuation as opposed to brackets ([ ]}. The inconvenience is compounded by this board's not allowing for editing after posting, or for a preview function.

So here is the punctuation-corrected version:


The Nature and Value of IP and VoIP

I wrote that "VoIP is a toy," and technoboy responded that "IP Communications is not a toy." This mistranslation is an invitation for me to clarify what I meant.

One thing that I did not mean -- and did not say -- was that "IP Communications is a toy." IP traffic between servers, i.e., "the Internet," is a useful thing. Essentially, it gives people easy access to libraries of information.

That's a lot different than VoIP as a standalone product, or even VoIP incorporated into all of the unnamed, unspecified wondrous new applications that no one can seem to describe. To denigrate VoIP is not to denigrate all of IP.

I have repeatedly stated here that I am highly skeptical of VoIP's potential for supplanting PCM, at least on the basis of any technological superiority. There are some regulatory issues favoring VoIP, not the least of which is the FCC's willingness to help the Regional Bells kill off what remains of the Telecom Act and restore their local wireline monopolies.

But if we separate those issues from the technology, there's simply no reason to migrate voice traffic from PCM to IP. It's not any cheaper; in fact, the opposite is probably the case when total costs are considered. IP's quality of service is always a fraction less than 1 of PCM's.

IP and the networks that carry it are by their very nature insecure relative to PCM. And the technology of real-time IP full-duplex comms remains quite immature, which means that any users will be laboratory rats for the propellerheads in the industry.

We all know, by the way, what propellerheads do when their products break. The blame the users. Nothing is ever their fault. Ever.

Finally, I'm not completely or unalterably dismissive of real-time, full-duplex IP. I have repeatedly stated my belief that IP videoconferencing is the killer app, not VoIP. If all the work on VoIP really winds up being used for IP videoconferencing, then it will almost certainly wind up being worthwhile.

And it sure won't be the first time that people found out that a particular technology was useful for something other than what they expected.


Intelligence in IP and PCM Networks

David Isenberg, a Bubble-era guru and shyster, wrote "The Dawn of the Stupid Network," in which he envisioned carriers as providers of dumb pipes, having no role other than bit transport.

Before even considering the thesis itself, all one really has to do is notice his Bubble hype, including the fulsome praise for the long-haul division of U.S. Worst/Qworst, which essentially dragged the company into scandal, corruption and bankruptcy.

I continue to be amazed by people's willingness to rally behind charlatans, which is exactly what I think Isenberg is. Had he written that future IP networks will be different than PCM -- and then detailed those differences -- I wouldn't be as harshly critical.

But he vastly overstated the case for the alleged simplicity of IP networking. He also misrepresented the nature of PCM networking. PCM networks feature considerable distribution of their intelligence, and IP networking will require the network operator to incorporate considerable intelligence.

A lot of semantic games have been played here, mainly with respect to whether the intelligence is at the core or the edge. This is unimportant relative to Isenberg's "stupid network" thesis, which argued that the network operator would simply maintain dumb pipes.

The fact is that, in a full-duplex, real-time IP network, the network provider will play a critical role that goes far, far beyond making sure that the glass doesn't break. Such IP networks will be highly deterministic, and as such they will be far more intelligent than the PCM network.

They will have to be more intelligent due to the variety of traffic they carry. PCM networks are purpose built for one application. Everything riding them gets the same level of QoS and security; IP networks couldn't be more different in this respect.

To argue, as Isenberg did, that the network operator will sort of wither away is foolish. Apparently, the California Disease is still with us. One of its elements is to never admit that you or your guru was just wrong.


Arrogance, Corruption and Truth

Communications technology is value-neutral. The only test is whether it works and at what cost. To make any particular moral argument for IP is foolish. Witness the religious wars between Ethernet and ATM. Who cares?

We know the companies with PCM networks. They are the phone carriers we've come to know and hate. When I write that PCM voice makes a lot more sense than VoIP, this is not any endorsement of the Regional Bells or the LD carriers.

Here are two basic truths about the PCM carriers:

1. All telecom carriers are evil.

2. The phone companies deliver a sack of flour and charge for cherry pie.

Here are three basic truths about the IP world:

1. The shit usually doesn't work, by which I mean it doesn't do anywhere near what the carriers or the vendors claim it will do, at a cost that makes any sense.

2. But for Moore's Law, the average vendor of hardware or software would the equivalent of a Florida land salesman of the 1920s. They tell any lie they can get away with, and because the products are so complicated they can get away with a lot of whoppers.

3. The IP world is characterized by a level of arrogance that would make the General Motors of the 1950s and 1960s blush. The PCM carriers would be exactly the same if they weren't so closely regulated.


Conclusion: Should We Just Shoot Ourselves, Then?

Hell no! One of the conceits of the IP people is that somehow their technologies will eliminate the tendency of human beings to lie, cheat and steal. You'd think that the experience of the Bubble, which is by a huge margin the biggest set of interconnected financial frauds the world has ever seen, would at least make these people a bit more humble.

But no, their naivete, greed and complete dishonesty continues unabated. I'm not quite sure why this is. Maybe there something in the water in California? I, for one, would like to see the IP crowd come down to earth.

But, if they don't, at least I will have the continuing pleasure of launching shit-missiles at them via the Light Reading board. In and of itself, that's enough reason to keep the knife away from my wrists.
technonerd 12/5/2012 | 2:04:11 AM
re: Mobile Skype: Quality Issues? [b][i]The Nature and Value of IP and VoIP[/i][/b]

I wrote that "VoIP is a toy," and technoboy responded that "IP Communications is not a toy." This mistranslation is an invitation for me to clarify what I meant.

One thing that I did [i]not[/i] mean -- and did not say -- was that "IP Communications is a toy." IP traffic between servers, i.e., "the Internet," is a useful thing. Essentially, it gives people easy access to libraries of information.

That's a lot different than VoIP as a standalone product, or even VoIP incorporated into all of the unnamed, unspecified wondrous new applications that no one can seem to describe. To denigrate VoIP is not to denigrate all of IP.

I have [i]repeatedly[/i] stated here that I am highly skeptical of VoIP's potential for supplanting PCM, at least on the basis of any technological superiority. There are some regulatory issues favoring VoIP, not the least of which is the FCC's willingness to help the Regional Bells kill off what remains of the Telecom Act and restore their local wireline monopolies.

But if we separate those issues from the technology, there's simply no reason to migrate voice traffic from PCM to IP. It's not any cheaper; in fact, the opposite is probably the case when total costs are considered. IP's quality of service is always a fraction less than 1 of PCM's.

IP and the networks that carry it are by their very nature insecure relative to PCM. And the technology of real-time IP full-duplex comms remains quite immature, which means that any users will be laboratory rats for the propellerheads in the industry.

We all know, by the way, what propellerheads do when their products break. The blame the users. Nothing is ever their fault. Ever.

Finally, I'm not completely or unalterably dismissive of real-time, full-duplex IP. I have repeatedly stated my belief that IP videoconferencing is the killer app, not VoIP. If all the work on VoIP really winds up being used for IP videoconferencing, then it will almost certainly wind up being worthwhile.

And it sure won't be the first time that people found out that a particular technology was useful for something other than what they expected.


[b][i]Intelligence in IP and PCM Networks[/i][/b]

David Isenberg, a Bubble-era guru and shyster, wrote "The Dawn of the Stupid Network," in which he envisioned carriers as providers of dumb pipes, having no role other than bit transport.

Before even considering the thesis itself, all one really has to do is notice his Bubble hype, including the fulsome praise for the long-haul division of U.S. Worst/Qworst, which essentially dragged the company into scandal, corruption and bankruptcy.

I continue to be amazed by people's willingness to rally behind charlatans, which is exactly what I think Isenberg is. Had he written that future IP networks will be different than PCM -- and then detailed those differences -- I wouldn't be as harshly critical.

But he vastly overstated the case for the alleged simplicity of IP networking. He also misrepresented the nature of PCM networking. PCM networks feature considerable distribution of their intelligence, and IP networking will require the network operator to incorporate considerable intelligence.

A lot of semantic games have been played here, mainly with respect to whether the intelligence is at the core or the edge. This is unimportant relative to Isenberg's "stupid network" thesis, which argued that the network operator would simply maintain dumb pipes.

The fact is that, in a full-duplex, real-time IP network, the network provider will play a critical role that goes far, far beyond making sure that the glass doesn't break. Such IP networks will be highly deterministic, and as such they will be far more intelligent than the PCM network.

They will [i]have to be more intelligent[/i] due to the variety of traffic they carry. PCM networks are purpose built for one application. Everything riding them gets the same level of QoS and security; IP networks couldn't be more different in this respect.

To argue, as Isenberg did, that the network operator will sort of wither away is foolish. Apparently, the California Disease is still with us. One of its elements is to never admit that you or your guru was just [i]wrong[/i].


[b][i]Arrogance, Corruption and Truth[/i][/b]

Communications technology is value-neutral. The only test is whether it works and at what cost. To make any particular moral argument for IP is foolish. Witness the religious wars between Ethernet and ATM. Who cares?

We know the companies with PCM networks. They are the phone carriers we've come to know and hate. When I write that PCM voice makes a lot more sense than VoIP, this is not any endorsement of the Regional Bells or the LD carriers.

Here are two basic truths about the PCM carriers:

1. All telecom carriers are evil.

2. The phone companies deliver a sack of flour and charge for cherry pie.

Here are three basic truths about the IP world:

1. The shit usually doesn't work, by which I mean it doesn't do anywhere near what the carriers or the vendors claim it will do, at a cost that makes any sense.

2. But for Moore's Law, the average vendor of hardware or software would the equivalent of a Florida land salesman of the 1920s. They tell any lie they can get away with, and because the products are so complicated they can get away with a lot of whoppers.

3. The IP world is characterized by a level of arrogance that would make the General Motors of the 1950s and 1960s blush. The PCM carriers would be exactly the same if they weren't so closely regulated.


[b][i]Conclusion: Should We Just Shoot Ourselves, Then?[/i][/b]

Hell no! One of the conceits of the IP people is that somehow their technologies will eliminate the tendency of human beings to lie, cheat and steal. You'd think that the experience of the Bubble, which is by a huge margin the biggest set of interconnected financial frauds the world has ever seen, would at least make these people a bit more humble.

But no, their naivete, greed and complete dishonesty continues unabated. I'm not quite sure why this is. Maybe there something in the water in California? I, for one, would like to see the IP crowd come down to earth.

But, if they don't, at least I will have the continuing pleasure of launching shit-missiles at them via the [i]Light Reading[/i] board. In and of itself, that's enough reason to keep the knife away from my wrists.
aswath 12/5/2012 | 2:04:17 AM
re: Mobile Skype: Quality Issues? Newer media gateways can the type of attack that you described. Also, please explain where these attacks eminate from and how they would get to the media gateways you describe. You should explain the overall architecture because I think all these statements are based on a fundamental mis-understanding of how these networks are being constructed.

I don't think we should interpret the description of the original scenario is suggesting that the gateways have a weakness. The question at hand is whether PSTN is vulnerable if VoIP is misbehaving. So we have to assume that the end-points and the gateways have been compromised. The argument concludes that even under this scenario, the integrity of PSTN is assured.
alchemy 12/5/2012 | 2:04:19 AM
re: Mobile Skype: Quality Issues? dljvjbsl writes:
Assuming that a mass attack such as talking about can be mounted, what will be the effect on the class 5.

When you attack a media gateway, you either attack signaling or media flow. The signaling is on a single IP address and a single port with a known protocol like MGCP. All you really need to do is hammer it with initialization messages at a fairly high rate. 1000/second will smoke most media gateways. When you attack the media stream, you're sending spurrious RTP packets that end up glitching individual phone calls and perhaps takes down part of the bank of codecs.

If you hit it with lots of traffic, you'll clog the IP pipe going into the box and take it down completely. These days, a carrier class media gateway has a GIG-E going to it so it takes quite a few zombie machines to kill it in this way.

In most architectures, you're actually talking to the Class 5 office via a separate signaling gateway. An attack on a media gateway can't harm the Class 5. Signaling gateways typically run SIGTRAN SS#7 over IP towards the IP network and a very small number of SS#7 over 56k links towards the SS#7 network. The Class 5 already knows how to handle SS7 congestion and can assert flow control by busying out CICs or other tried & true mechanisms. It's easy to take out a signaling gateway but it's unlikely you'll be able to take down anything in the PSTN network.
technoboy 12/5/2012 | 2:04:19 AM
re: Mobile Skype: Quality Issues? alchemy:When you attack a media gateway, you either attack signaling or media flow. The signaling is on a single IP address and a single port with a known protocol like MGCP. All you really need to do is hammer it with initialization messages at a fairly high rate. 1000/second will smoke most media gateways. When you attack the media stream, you're sending spurrious RTP packets that end up glitching individual phone calls and perhaps takes down part of the bank of codecs.

If you hit it with lots of traffic, you'll clog the IP pipe going into the box and take it down completely. These days, a carrier class media gateway has a GIG-E going to it so it takes quite a few zombie machines to kill it in this way.

Tell me where this has occurred and also you are referring to a media gateway which using MGCP or perhaps megaco. This was popular a few years ago when carriers were moving some traffic from the class 5 to media gateway and then sending it across their backbone. That is fundamentally diffeent than what is envisioned today. Newer media gateways can the type of attack that you described. Also, please explain where these attacks eminate from and how they would get to the media gateways you describe. You should explain the overall architecture because I think all these statements are based on a fundamental mis-understanding of how these networks are being constructed.
technoboy 12/5/2012 | 2:04:20 AM
re: Mobile Skype: Quality Issues? dreamer101 wrote:Hi technoboy,
The title of the subject line is "VoIP vs. POTS". So, VoIP in the case of enterprise using IP PBX and inter-connecting them via private IP network is not the same as as replacing POTS.

To replace POTS, you will end up something like Packet8 or Vonage where your VoIP phone is connected to the VoIP media gateway via Internet. What I said will not be true only for some VoIP service provide like MSO where they connect to the VoIP phone via private IP network directly without Internet.

Hi Dreamer101,
Well this message thread has gone off in several directions and technonerd has mentioned the enterprise PBX and VOIP in a couple of them hence my response to that the particular statement. Regarding the statement that only MSO' are building private IP networks, that I can tell you is false. All of the service providers that I am aware of are building their own networks and adapting the existing traffic and interfaces to run on one network which is IP. It is not logical to think these companies are going to run their service through the publicn internet. They may leverage the public internet for certain components such as mobil or teleworker applications as part of an overall service offering.
dljvjbsl 12/5/2012 | 2:04:20 AM
re: Mobile Skype: Quality Issues?
Here we agree up until the physical point where you have to connect to a POTS phone (which is generally connected via GR-303 interfaces to a class 5 switch). The amount of IP-based phones is substantially less than 1% of the in-service phones today. I find your defenses based on a currently almost non-existent network architecture disturbing. VoIP is the 'technological dislocation' that hasn't reached anywhere near critical mass to be considered over or even 'accepted'.


If this is the scenario that concerns you, then we must consider what sort of gateway is connecting the VoIP to the PSTN network. As others have pointed out before, this will be through a set of trunks. Assuming that a mass attack such as talking about can be mounted, what will be the effect on the class 5. The trunks between the VoIP gateway and the class 5 will be quickly exhausted and the class 5 will carry on as before. The trunking gateway would seem to be the choke point. Just not enough signalling traffic can pass through the trunks to affect the call processing on the class 5.
sgan201 12/5/2012 | 2:04:22 AM
re: Mobile Skype: Quality Issues? Hi,
Anyhow, I think this discussion thread is highly irrelevant. VoIP (as a replcement for POTS) has a very limited and negligible market for places where you can get POTS and cellular services.

It is more interesting to discuss how and if possible to build a public VoIP (as a replacement for POTS) network where the POTS and cellular infracstructure does not exist. And, that is where the growth is..

Dreamer
technoboy 12/5/2012 | 2:04:23 AM
re: Mobile Skype: Quality Issues? All that has to happen to kill VoIP is a single class 5 being isolated for (insert some period of time) before government regulations, lost-business lawsuits, carriers cancelling orders, etc. due to the weakest link in that chain being compromised.
Absolutely the case! Yes, there are a few IT departments out there messing around with VoIP but I would be surprised if any of them have done away with the POTS lines. Most businesses know that if the phones go down they're not in business. VoIP is a a toy.

IP Communications is not a toy. The way people on this board interchange the facts with their own versions of reality is quite humorous. Quantify a vew departments playing with VOIP. Is that number 3, 30, 300, 3000. How many commercial customers do you really think are using the technology today in a production network. Check your facts because the enterprise is embracing the technology much faster than you realize. You can argue the reasons why or if it makes sense but you cant argue whether or not it is happening. It is happening today and the technology will continue to grow in the enterprise space. If the argumnet is that no one will use VOIP because it is too susceptible to security issues than I would pose the same issues that affect IP communuications affect misssion critical applications for any enterprise. Funny thing is that this has not prevented companies from deploying those applications and using them despite all of the dire consequences of how worms and viruses will bring these networks to there knees. we might as well move back to the days of green screens if that is the view.

sgan201 12/5/2012 | 2:04:23 AM
re: Mobile Skype: Quality Issues? Hi technoboy,
The title of the subject line is "VoIP vs. POTS". So, VoIP in the case of enterprise using IP PBX and inter-connecting them via private IP network is not the same as as replacing POTS.

To replace POTS, you will end up something like Packet8 or Vonage where your VoIP phone is connected to the VoIP media gateway via Internet. What I said will not be true only for some VoIP service provide like MSO where they connect to the VoIP phone via private IP network directly without Internet..

Dreamer
technoboy 12/5/2012 | 2:04:24 AM
re: Mobile Skype: Quality Issues? I think some of these statements are just all over the map. Talk about mass DDOS attacks and viruses taking down the internet are all possible but I dont really think they are accurate in context of IP Communication networks. Service providers are not relying solely on the internet to provide their communications networks. They are building their own IP backbones and slowly extending them out to the edge and then to the access layer. Im not talking about companies like 8x8 or Vonage Im talking about RBOCS and IXC's. These networks may connect to the internet ( a rather loose term in my opinion) but the backbones are truly their own.
PO 12/5/2012 | 2:04:24 AM
re: Mobile Skype: Quality Issues? Fortunately, the experience of 9/11 inidcates that the Intrnet is robust in the face of mass calling events.

I must disagree most emphatically on this point. The experience on the Internet is that resources may exhaust at any of a number of points along any of a number of paths, and requires effort to manage just as on the TDM network.

We've been almost fortunate that a few companies, such as Akamai, dominate the actual traffic sources, so we don't notice many significant events: they are instead absorbed into massive shared server pools. This dominance, however, is not a feature of the network but rather an artifact of one company's corporate successes.

Setting mirrors aside, mass calling events on the Internet may exhaust bandwidth, CPU, memory, or other resources quite readily, and preparations must be taken in line with the likelihood of such events.
technonerd 12/5/2012 | 2:04:25 AM
re: Mobile Skype: Quality Issues? If they found out that an outage of similar scale was caused by a single or group of hackers because of access enabled by adoption of VoIP technology there will be hell to pay; especially if someone dies as a result.
Hey, Steve get with the program! You wanna make an omelette you've gotta break some eggs. People are gonna die. Get used to it!


If any disruption is caused by hackers connected to a terrorist organization the government itself will legislate the death of VoIP gateways to the POTS network.
Unless the terrorist organization is a big campaign contributor, in which case they'll send the chairman of the FCC out to say that there wasn't a specific enough warning.


I guess the part that makes me so sure that this type of thing will certainly occur is the arrogance of those who say it cannot.
Of course it will happen. And if the right Wall Street bankers and the right VCs who have bought the right politicians are behind it, the whole thing will be swept under the rug.


Now I have already admitted that I am no genius and I have been humbled because I used to think that I was pretty good. Fortunately my oversights were in networks that were totally isolated from external, malicious geniuses. VoIP has opened Pandora's box in ways that we cannot yet imagine.
This is the problem, Steve. It will keep you from ever getting enough stock options to be rich. The first thing you must do is develop the skills to shift the blame on people lower than yourself.
technonerd 12/5/2012 | 2:04:25 AM
re: Mobile Skype: Quality Issues? Hardened VoIP services use specialized boxes to classify VoIP signaling and media flows and limit the damage of distributed denial of service attacks. You'll see these boxes called 'Session Controllers' or 'Border Proxies'. They run at wire speed and only let through packets that they're expecting and limit the rate of packets so nobody can damage the network.

Hmm. That wouldn't be an example of putting intelligence into the "Stupid Network," would it? No, of course not! Perish the thought!
stephenpcooke 12/5/2012 | 2:04:26 AM
re: Mobile Skype: Quality Issues? Now, who brought in the Internet traffic being affected...? I thought that we had agreed that we were talking about the TOLL or POTS network?!!!

The issue on that has been decided. The curent toll network is being replaced by an IP network based on MPLS. This debate seems to be over until the next technological dislocation takes place.

This will be completed when...? Until that time only God can ensure that an event such as I have described will not happen.

VoIP traffic in any event will not be carried across any remenant of the TDM toll network. it will be carried on the IP network simply for cost reasons.

Here we agree up until the physical point where you have to connect to a POTS phone (which is generally connected via GR-303 interfaces to a class 5 switch). The amount of IP-based phones is substantially less than 1% of the in-service phones today. I find your defenses based on a currently almost non-existent network architecture disturbing. VoIP is the 'technological dislocation' that hasn't reached anywhere near critical mass to be considered over or even 'accepted'.

Mass calling events of the scale that is now possible due to VoIP-enabled home computers infected with viruses and/or worms may only have been seen with 9/11. The public was willing to forgive the disruptions caused in that case for obvious reasons. If they found out that an outage of similar scale was caused by a single or group of hackers because of access enabled by adoption of VoIP technology there will be hell to pay; especially if someone dies as a result. If any disruption is caused by hackers connected to a terrorist organization the government itself will legislate the death of VoIP gateways to the POTS network.

I guess the part that makes me so sure that this type of thing will certainly occur is the arrogance of those who say it cannot. How many OS designers intentionally left holes for viruses to take root? How many people thought the Titanic was 'unsinkable'? How many outages did I myself have to debug in a product that I had significant technical input into? I documented every product shortcoming that I could conceive of and there were still odd ones that got through. Now I have already admitted that I am no genius and I have been humbled because I used to think that I was pretty good. Fortunately my oversights were in networks that were totally isolated from external, malicious geniuses. VoIP has opened Pandora's box in ways that we cannot yet imagine.
alchemy 12/5/2012 | 2:04:26 AM
re: Mobile Skype: Quality Issues? Hardened VoIP services use specialized boxes to classify VoIP signaling and media flows and limit the damage of distributed denial of service attacks. You'll see these boxes called 'Session Controllers' or 'Border Proxies'. They run at wire speed and only let through packets that they're expecting and limit the rate of packets so nobody can damage the network.

With a bare-naked SIP Proxy (typically a Sun Netra), it's quite easy to mount a denial of service attack that will kill it. All I have to do is nail it with 5000 REGISTER messages per second from a bunch of zombie machines and I've created a doorstop.
technonerd 12/5/2012 | 2:04:29 AM
re: Mobile Skype: Quality Issues? "Congestion occurred on the periphery" If you mean that thousands of people in areas in the general NY vicinity couldn't get dialtone, including people who were not on the Wall Street switches that were destroyed, then you are correct.
And if I'm not mistaken, e-mails weren't getting through to a lot of places because their access lines were down. I suspect some ISPs had trouble, too.

How many home computers still use Windows98 without any built in firewall? How many dial-up users have a firewall at all? How many high-speed users even know what a firewall is let alone its purpose?
Not only that, but the security game with IP is a dog-chasing-its-tail kind of thing and always will be. IP was not designed with security in mind. All security systems are bolt-ons; it is always a spy-vs-spy game on the Internet and it always will be.


All that has to happen to kill VoIP is a single class 5 being isolated for (insert some period of time) before government regulations, lost-business lawsuits, carriers cancelling orders, etc. due to the weakest link in that chain being compromised.
Absolutely the case! Yes, there are a few IT departments out there messing around with VoIP but I would be surprised if any of them have done away with the POTS lines. Most businesses know that if the phones go down they're not in business. VoIP is a a toy.
technonerd 12/5/2012 | 2:04:29 AM
re: Mobile Skype: Quality Issues? Mass calling events are a serious concern for the traditional PSTN. It does not take a disaster on the scale of 9/11 to create them. Radio station contests can trigger them as well. They are a very serious issue - in this I agree with you.

Fortunately, the experience of 9/11 inidcates that the Intrnet is robust in the face of mass calling events. This is additionally fortunate since the toll voice network is being converted to a new network based on IP technology.


The only Internet "mass calling event" that I've heard of is when the whole thing slowed down because of the Victoria's Secret thing on one Super Bowl Sunday.
dljvjbsl 12/5/2012 | 2:04:30 AM
re: Mobile Skype: Quality Issues?
Now, who brought in the Internet traffic being affected...? I thought that we had agreed that we were talking about the TOLL or POTS network?!!!


The issue on that has been decided. The curent toll network is being replaced by an IP network based on MPLS. This debate seems to be over until the next technological dislocation takes place.

VoIP traffic in any event will not be carried across any remenant of the TDM toll network. it will be carried on the IP network simply for cost reasons.


"Congestion occurred on the periphery" If you mean that thousands of people in areas in the general NY vicinity couldn't get dialtone, including people who were not on the Wall Street switches that were destroyed, then you are correct


Mass calling events are a serious concern for the traditional PSTN. It does not take a disaster on the scale of 9/11 to create them. Radio station contests can trigger them as well. They are a very serious issue - in this I agree with you.

Fortunately, the experience of 9/11 inidcates that the Intrnet is robust in the face of mass calling events. This is additionally fortunate since the toll voice network is being converted to a new network based on IP technology.


You are assuming that only a single proxy is being utilized here. In my original post I described the fact that my firewall had intercepted illegal login attempts to my home computer from hackers or "zombies" worldwide. The question gets more and more valid as the number of proxies (ie: the amount of VoIP equipment deployed) increases. DDoS attacks happen on a HUGE scale.


I would remind you that the VoiP network is indeed a network. Trust relationships will be created within the network and this trust will be enforced by cryptographic techniques. Proxies will be accepting traffic only from trusted sources.


I give you VoIP guys credit, you are trying hard but I have heard nothing so far that totally convinces me that what I have described cannot happen.


The DSC SS7 overload bug shut down the network in the Eastern US so the problem has already happened in the PSTN. VoIP networks have not created this risk.

One can only trust that problems as they occur will be dealt with. My own assumption is that the demonstrated robustness of Internet switching makes events such as these more unlikely.
aswath 12/5/2012 | 2:04:31 AM
re: Mobile Skype: Quality Issues? Oh by the way that sewage pipe goes to a sewage treatment plant where the sewage is treated and discharge into a stream ... Guess what you dont know the difference do you?

The analogous treatment plant for VoIP/PSTN interworking is gateway and the filter is protocol termination. Unlike water treatment filter, this filter is tight.
technoboy 12/5/2012 | 2:04:31 AM
re: Mobile Skype: Quality Issues? dreamer101: < I have two set of pipes in my house, water and sewage. Why would I want to combine them?"

1) It is convergence..

2) Because Cisco told you that it is the right thing to do..

You give way too much credit to cisco if you think they have that level of influence on how our telecom infrastructure will evolve.

Oh by the way that sewage pipe goes to a sewage treatment plant where the sewage is treated and discharge into a stream which deposits water back into the groud as well as carrying it to a lake, reservoir or ocean or which partially goes back into the air all of which in one form or another comes back to your house eventually as tap water. Guess what you dont know the difference do you?
sgan201 12/5/2012 | 2:04:32 AM
re: Mobile Skype: Quality Issues? Hi stephenpcooke,
< I have two set of pipes in my house, water and sewage. Why would I want to combine them?"

1) It is convergence..

2) Because Cisco told you that it is the right thing to do..

I could not resist commenting on this one..

Dreamer
stephenpcooke 12/5/2012 | 2:04:34 AM
re: Mobile Skype: Quality Issues? dljvjbsl wrote:

The issue concerns the possibility of hijacking enough of those clients so that they could consume a signgicantg potion of the procesing power of the toll network.

That will do to start...

Firstly I do not see that VoIP calls will be traversing the traditional toll network. Even calls that are routed to POTS phones will be kept IP for as long as possible. Even in the mass calling event caused by 9/11, core Internet experienced no problem in absorbing the extra traffic. Congestion occurred on the periphery.

Now, who brought in the Internet traffic being affected...? I thought that we had agreed that we were talking about the TOLL or POTS network?!!!

"Congestion occurred on the periphery" If you mean that thousands of people in areas in the general NY vicinity couldn't get dialtone, including people who were not on the Wall Street switches that were destroyed, then you are correct.

Mass calling events would seem to be choked at the local proxy.

You are assuming that only a single proxy is being utilized here. In my original post I described the fact that my firewall had intercepted illegal login attempts to my home computer from hackers or "zombies" worldwide. The question gets more and more valid as the number of proxies (ie: the amount of VoIP equipment deployed) increases. DDoS attacks happen on a HUGE scale. There are literally hundreds of BILLIONS of access attempts on the targetted servers. These attacks are generally carried out by thousands and thousands of infected home computers and poorly protected corporate network servers. How many home computers still use Windows98 without any built in firewall? How many dial-up users have a firewall at all? How many high-speed users even know what a firewall is let alone its purpose?

Thirdly, the size of the network would seem to mitigate this as well. As hte 9/11 experience showed even with a mass calling event of unprecedented proportion the Internet was not overloaded. The difficulty in recuiting enough zombies to make an appreciable difference in Internet traffic would seem to argue strongly against the practical import of the threat that you describe.

So, I have dealt with the last part of your comment above. The first part has little to do with my concern which is the safety of the POTS network, not the Internet.

Let me re-state and please know that I am not trying to say that the 'best and brightest' people working on this have missed the boat here, I'm sure they are working hard to prevent this type of thing happening because it would be the death-knell for VoIP. What I am concerned about is the POTS network, period. There are many potential difficulties already resident in the POTS network but there are none, that we know of, that have the potential disruptive power of what I have described.

As long as 'virtual', non-human-initiated phone calls can be generated by (put your favorite Internet-based nasty here, making sure that you include viruses, adware, spyware, worms, etc.)-infected computers, the POTS network is vulnerable. And lets be clear, the POTS network is a huge chain that is built on things like bandwidth, signalling, statistical loading, wavelengths, wireless terminals (which, if Internet-enabled could also be a source of these problems), etc. All that has to happen to kill VoIP is a single class 5 being isolated for (insert some period of time) before government regulations, lost-business lawsuits, carriers cancelling orders, etc. due to the weakest link in that chain being compromised.

I give you VoIP guys credit, you are trying hard but I have heard nothing so far that totally convinces me that what I have described cannot happen.

Comment heard at an industry conference: "I have two sets of pipes in my house, water and sewage. Why would I want to combine them?"
technoboy 12/5/2012 | 2:04:35 AM
re: Mobile Skype: Quality Issues? DVJBSL wrote:
Calls to the will not be accepted randomly but will only be accepted from trusted servers. Calls from residential gateways will be handled by proxies which will then route them. One of the major issues with SPAM is the ease with which identities may be spoofed. This is not the case in SIP. There has been a lot of work on authentication and authorization.

Technonerd:Hmm, doesn't much sound like the brilliant Mr. Isenberg's "stupid network," does it? Seems like the carriers will need to put quite a bit of intelligence into their dumb pipes.

I think this exchange illustrates an important point. You have to differentiate between a network and the applications and other devices that communicate across this network. When I authenticate to a SIP proxy the intelligence that allows this to happen and for the interaction to occur is between my end device ( IP phone, or soft phone etc...) and the SIP proxy. The routers and switches are not providing any higher level authentication or interaction in this instance. It would probably be helpful if we all agree on what intelligent really means. I had stated in other posts that the networks being constructed today consist of an access layer that can adapt many traffic types into an IP format. The traffic types are clissified at the edge of the network (QOS and Security). The traffic is then mapped into an MPLS path. The core of these networks will consist of IP over an optically meshed layer.

If you defines intelligence as looking at layer 3 and above for the entire path of the network than I would tell you that it is not likely to happen in my view. There is no real reason to look at the application layer in the core of a network nor is it useful to look at implementing security at the core of a network. This is best handled at the edge where networks exchange information with other networks.
technoboy 12/5/2012 | 2:04:35 AM
re: Mobile Skype: Quality Issues? technonerd wrote: If someone wants to host a video conference, a request will go out to the carrier for bandwidth and QoS adequate to handle it. The bits themselves will pass through equipment that can read all seven layers of that stack, and charge for the capability. And if there's a wiretap happening, the equipment will facilitate that, too.

Just curious what service provider do you talking about this type of network. I think this type of capability is far into the future. Perhaps another generation past the networks that are currently being built.

The networks that are currently envisioned do not look to provide granular control or QOS from the edge to the core. The acces layer adapts the various speeds and feeds and traffic types to an IP format. The edge of the network classifies this traffic and maps to an MPLS path. The core is nothing more than large pipes over an optical mesh running at very high speed. This means that core networks are not running QOS marking or security controls and definitely not looking into the application layer because this has been handled at the edge and is better suited to be handled at the edge. Even provisioning of routing protocols can be done at the edge. So the network core is dumb in the sense that it is built to be very redundant and have many high speed links. Traffic marked at the edge passes through the core depending on its assigned path. Meshing as opposed to SONET allows for more survivability and more paths through the network.
technonerd 12/5/2012 | 2:04:36 AM
re: Mobile Skype: Quality Issues? One question -- how will the central operators understand what the application is? Applications will be defined across multiple carrier and enterprise networks. Applications will be defined at the periphery.

If someone wants to host a video conference, a request will go out to the carrier for bandwidth and QoS adequate to handle it. The bits themselves will pass through equipment that can read all seven layers of that stack, and charge for the capability. And if there's a wiretap happening, the equipment will facilitate that, too.


The stupid network does not mean that there can be no processing in the central network. It means that the central network will be controlled by the periphery.
Just as you have now on the PSTN, albeit with a single application. Equipment at the edge makes requests and the network fulfills them. In any case, I still haven't had any examples of these great new applications that you're envisioning.


Why is so difficult for you to accept that other people cna have good ideas. Companies such as IBM, Microsoft, Sun ... are all developing products using these ideas. Are they just "bubble-era blowhards?"
Isenberg is not IBM or Microsoft of Sun. He's one of a coterie of bubble people who got way too much ink. I guess what bugs me is that his crew served as the catalyst for untold billions in stupid speculation, yet somehow we still have people kissing their asses. It's just ridiculous.
technonerd 12/5/2012 | 2:04:40 AM
re: Mobile Skype: Quality Issues? The stupid network idea works at the level of call processing not transmission (i.e. PCM). The ability of IP/packet networks to provide for the dynamic creation of connections by the end points creates advantages that the traditional PSTN just cannot cope with. End points can set up control and data connections at will.

With the 'stupid network', network operation is slaved to the intelligence at the periphery. This allows a close convergence between voice and data applications at the periphery and hence more customer value.


You seem to continue to ignore a vital truth. Both a PCM-based and an IP-based network operator must enforce and insure QoS. In doing so, the network operator provides "intelligence." If you wish to define the "stupid network" as the Layer 0 facilities between boxes, well then yes IP networks can be stupid networks. And so can PCM networks be "stupid."

But that's foolish. What really happens is that the network operator will build various pieces of intelligence into his network, be it PCM or IP. For example, if someone wants to hold an IP video conference on an IP network, they're going to need a certain level of QoS to do it. They'll also need a certain level of security.

The endpoints will certainly be able to request these levels, but it will be the network that will provide and enforce them. Oh, and bill for them. You think this is somehow going to be free? You know, like the free electricity from the nuclear power plant?

Will the endpoints in an IP network be theoretically able to define more services than a PCM network can? Let's hope so, given that packets can represent not just voice, but video and data as well. But does the mean that they will somehow replace PCM nets for voice, which has been the leitmotif of all these discussions?

I say not because of any technology issue. There are regulatory considerations responsible for pushing IP in general and VoIP in particular. International toll bypass, domestic access charge bypass and the FCC's willingness to allow the Regional Bells to define VoIP as an information service, thereby putting the final nail into the coffin of the Telecom Act.

But on the technology and applications side, there simply aren't any real examples that people can show of user-defined VoIP applications that matter. "Technoboy" came up with one, but it was pretty obscure and fairly weak to be quite honest. At the moment, I think what we have here is a distinction without a difference.


As Isenberg and many others point out this is not a choice that the carriers can choose to make it is a basic technological imperative that is driven by the self-interest of users. The carriers with their class 5s and centralized control will become irrelevant.

Class 5s are going to be handling calls in even the most advanced networks for a very, very long time. Many more decades. But even if they're replaced by big routers, the network operators are still going to be central. I mean, look at China. They government (with the help of Cisco, et al.) taps into their packet networks to the degree that they can identify and isolate those who should be tortured and killed for transmitting the wrong packets. This is a stupid network? Don't think so.

In the coming era, the central operators will have more control not less. This is because their equipment will examine all seven layers of the stack plus the content. These capabilities will be on the network side of the demarc.

Face it, Isbenberg was just one more bubble-era blowhard who got it wrong. He's been wrong about general issues and wrong about specific companies. He's been right about virtually nothing.
dljvjbsl 12/5/2012 | 2:04:40 AM
re: Mobile Skype: Quality Issues?
In the coming era, the central operators will have more control not less. This is because their equipment will examine all seven layers of the stack plus the content. These capabilities will be on the network side of the demarc.


One question -- how will the central operators understand what the application is? Applications will be defined across multiple carrier and enterprise networks. Applications will be defined at the periphery.

There is much good technology being developed to accomlish this. Policy-based systems, directory-enabled netowrks, active networks ... are all being developed so that peripheral applications can control the behavior of network elements. The stupid network does not mean that there can be no processing in the central network. It means that the central network will be controlled by the periphery.


Isbenberg was just one more bubble-era blowhard who got it wrong. He's been wrong about general issues and wrong about specific companies. He's been right about virtually nothing


Why is so difficult for you to accept that other people cna have good ideas. Companies such as IBM, Microsoft, Sun ... are all developing products using these ideas. Are they just "bubble-era blowhards?"
technonerd 12/5/2012 | 2:04:40 AM
re: Mobile Skype: Quality Issues? I think the distinction that you are trying to make is between distributed packet switching and centralized TDM switching.

So, do you think all TDM switching is centralized? If so, what about a PBX? Seems to me that this is just as "distributed" as any router I've ever heard of.
dljvjbsl 12/5/2012 | 2:04:42 AM
re: Mobile Skype: Quality Issues?
The big carriers like ATT, MCI, et. al., use PCM in most places but IP in others and will keep doing so.


The PCM that you are talking about is usually 64Kb/s mu law or A law. Both of these codecs (and others) may be used by VoIP.

I think the distinction that you are trying to make is between distributed packet switching and centralized TDM switching.
dljvjbsl 12/5/2012 | 2:04:42 AM
re: Mobile Skype: Quality Issues?
Sounds like VoIP networking will be a lot more deterministic than what's been associated with the best-efforts service between servers that people have labeled "the Internet." Why, of course, because if you want to run full duplex in realtime, then you must be more deterministic.

Two implications fall out from this. One is that the network has to be intelligent, and second is that there ain't no way it's going to be "free" as so many propellerheads seem to think.


It is very difficult to know where to start here. I suppose that I could say that there is a distinction between transmission and call processing that you do not seem to comprehend.

The stupid network idea works at the level of call processing not transmission (i.e. PCM). The ability of IP/packet networks to provide for the dynamic creation of connections by the end points creates advantages that the traditional PSTN just cannot cope with. End points can set up control and data connections at will.

With the 'stupid network', network operation is slaved to the intelligence at the periphery. This allows a close convergence between voice and data applications at the periphery and hence more customer value.

As Isenberg and many others point out this is not a choice that the carriers can choose to make it is a basic technological imperative that is driven by the self-interest of users. The carriers with their class 5s and centralized control will become irrelevant.


Frankly, security in the IP world is a bolt-on as opposed to it being part of PCM from the get-go. As a result, using IP for real-time, mission-critical traffic will [i]ALWAYS[/i] carry a higher security risk than using PCM.


There is something of value in this observation. However I do not think that the value was intended.

The value that the stupid network brings relies on the easy connectivity supplied by packet (i.e. IP) switching. End points can set up connections for both data (including voice) and control at their ease. This allows dynamically created applications to use the network. Control channels are created at will.

Unfortunately, this type of in-band (end-to-end) control exposes the network to malicious control attempts from the periphery. The PSTN was not designed from the get go to handle this type of risk. Blue box (SF 2600hz trunk) and red box (coin phone signals) frauds resulted directly from the use of in-band signaling by the PSTN. The PSTN, to handle these frauds, moved to out-band signaling in which the periphery did not have access to the network control channels.

VoIP just like the PSTN is vulnerable to frauds created by the use of in-band signaling. Thus the designers of VoIP protocols have worked very hard on the issues of authentication and authorization so that the VoIP control channels are isolated and protected by cryptographic techniques. Malicious end points will be faced with difficult cryptographic barriers if they attempt to damage the network. VoIP uses out-band signaling channels.

Hence VoIP designers have learned from the vulnerability of the PSTN to in-band frauds by isolating their control channels from the get go contrary to what you have asserted.
technonerd 12/5/2012 | 2:04:43 AM
re: Mobile Skype: Quality Issues? Even in the mass calling event caused by 9/11, core Internet experienced no problem in absorbing the extra traffic. Congestion occurred on the periphery.
The same is true of POTS. Any problems were just as much at the "periphery" as you define it. Why? Because there is no physical "core Internet." The Internet is a logical entity.


Secondly, the VoIP network is just that a network.
There is no such thing as "the VoIP network." VoIP networks are not physical entities, they're logical entities. The big carriers like ATT, MCI, et. al., use PCM in most places but IP in others and will keep doing so.


Calls to the will not be accepted randomly but will only be accepted from trusted servers. Calls from residential gateways will be handled by proxies which will then route them. One of the major issues with SPAM is the ease with which identities may be spoofed. This is not the case in SIP. There has been a lot of work on authentication and authorization.
Hmm, doesn't much sound like the brilliant Mr. Isenberg's "stupid network," does it? Seems like the carriers will need to put quite a bit of intelligence into their dumb pipes.


As others have indicated previously, during mass calling events in the PSTN, originations from certain classes of DNs (directory numbers) may be deprecated in preference to others. For example home phones will wait for dial tone while hospitals will be given priority. With SIP ths could be extended to include DNs that are originating a great deal of traffic. Their originations could simply be ignored. Mass calling events would seem to be choked at the local proxy.
Sounds like VoIP networking will be a lot more deterministic than what's been associated with the best-efforts service between servers that people have labeled "the Internet." Why, of course, because if you want to run full duplex in realtime, then you must be more deterministic.

Two implications fall out from this. One is that the network has to be intelligent, and second is that there ain't no way it's going to be "free" as so many propellerheads seem to think.


The difficulty in recuiting enough zombies to make an appreciable difference in Internet traffic would seem to argue strongly against the practical import of the threat that you describe.
Oh, I'd say there are plenty of technozombies available in California for the picking. And, at this point, one technozombie can do more damage in the IP world than in the PCM world because the IP world hasn't paid all that much attention to security.

Frankly, security in the IP world is a bolt-on as opposed to it being part of PCM from the get-go. As a result, using IP for real-time, mission-critical traffic will [i]ALWAYS[/i] carry a higher security risk than using PCM.
dljvjbsl 12/5/2012 | 2:04:43 AM
re: Mobile Skype: Quality Issues?
perhaps you can describe this limitation? I'm no genius but if it was my goal to bring down a bunch of toll switches across the US I'd have all the domestic gateways max'ed out on every class 5 making long distance calls to POTS lines, from every country with a gateway.


The issue that I see that you are bringing up is the use of intelligent clients in the periphery. The issue concerns the possibility of hijacking enough of those clients so that they could consume a signgicantg potion of the procesing power of the toll network.

Firstly I do not see that VoIP calls will be traversing the traditional toll network. Even calls that are routed to POTS phones will be kept IP for as long as possible. Even in the mass calling event caused by 9/11, core Internet experienced no problem in absorbing the extra traffic. Congestion occurred on the periphery.

Secondly, the VoIP network is just that a network. Calls to the will not be accepted randomly but will only be accepted from trusted servers. Calls from residential gateways will be handled by proxies which will then route them. One of the major issues with SPAM is the ease with which identities may be spoofed. This is not the case in SIP. There has been a lot of work on authentication and authorization. As others have indicated previously, during mass calling events in the PSTN, originations from certain classes of DNs (directory numbers) may be deprecated in preference to others. For example home phones will wait for dial tone while hospitals will be given priority. With SIP ths could be extended to include DNs that are originating a great deal of traffic. Their originations could simply be ignored. Mass calling events would seem to be choked at the local proxy.

Thirdly, the size of the network would seem to mitigate this as well. As hte 9/11 experience showed even with a mass calling event of unprecedented proportion the Internet was not overloaded. The difficulty in recuiting enough zombies to make an appreciable difference in Internet traffic would seem to argue strongly against the practical import of the threat that you describe.
technonerd 12/5/2012 | 2:04:49 AM
re: Mobile Skype: Quality Issues? Applications will be defined at the periphery. VoIP proxies are at the periphery and are not owned by the carrier. This solves problems that cannot be solved by the AIN because it trys to fit a one-size-fits-all soultion to all users.

As long as the carrier enforces QoS, the "stupid network" thesis remains, well, stupid.
technonerd 12/5/2012 | 2:04:49 AM
re: Mobile Skype: Quality Issues? 802.11 handoff is available from at least one company thatI know of and I assume more. I really do not think that handoff is all that important for nomadic users. One sits at a meeting room or restarant table to make connections to the network

Handoff technology is irrelevant. The frequencies used by 802.11 are so high that you'd need way too many cellsites to provide the sort of coverage that would make mobility workkable. I suspect handoff technology is designed for very limited uses, i.e., within a building or maybe on a campus.

This limits 802.11 to portable but not mobile use. And if it's not mobile, then "Wi-FI" isn't going to be a competitor to cellular voice worthy of mention. Besides that, even in a portable environment you're going to require an ISP with voice-over enablement, i.e., a VoIP carrier.

That service simply is not free other than to a hobbyist here and there who can put up with the problems. Once you use a VoIP carrier it will charge you for the service, in which case we run into a cost comparison with cellular voice. And, as we all know, cellular voice is cheap enough that the bills don't matter -- especially for corporate users.

Thus, the idea that "Wi-Fi" voice is somehow going to be competitive with cellular voice is just foolish. It's a propellerhead fantasy. Who knows, if the tech stocks stay as overheated as they are then maybe someone will float in IPO or two and we'll hear all about the wonders of VoIP-over-WiFi in Light Reading and other objective trade journals like Wired, not to mention reports from those other bastions of integrity, Morgan Stanley, Merrill Lynch, et. al.
dljvjbsl 12/5/2012 | 2:04:51 AM
re: Mobile Skype: Quality Issues?
Interesting discussion. Doesn't seem like the network will be a mere collection of stupid pipes, does it? Perhaps Mr. Isenberg was a bit off base? Nah.


You really do not understand what Isenberg, Luckey and all the rest are saying.

Applications will be defined at the periphery. VoIP proxies are at the periphery and are not owned by the carrier. This solves problems that cannot be solved by the AIN because it trys to fit a one-size-fits-all soultion to all users.
dljvjbsl 12/5/2012 | 2:04:52 AM
re: Mobile Skype: Quality Issues?
This thread began with a discussion of using Wi-Fi for VoIP service. Unless you can do handoffs, it's foolish to imagine that VoIP over Wi-Fi is going to be anything other than a geek toy.


Not that I think that this is a very important topic, but 802.11 handoff is available from at least one company thatI know of and I assume more. I really do not think that jandoff is all that important for nomadic users. One sits at a meeting room or restarant table to make connections to the network.
dljvjbsl 12/5/2012 | 2:04:52 AM
re: Mobile Skype: Quality Issues? referring to WiFi:

It's not popular within businesses. It has a measure of popular with the IT geeks. Not with anyone else.


This reminds me of the Yogi Berra observation:

"Nobody goes there anymore; it's too crowded."
stephenpcooke 12/5/2012 | 2:04:53 AM
re: Mobile Skype: Quality Issues? PO wrote:
You'll want to get a copy of logresolve, which will do a reverse lookup from the IP addresses in a whole log file.

My firewall, bizguardian, does not have this feature but I was wondering if anyone else knows if it exists: I would want to take the log file, do a reverse lookup on all the IP addresses and send the list of pertinent offenders to their individual ISPs. Granted this may be a waste of time as many ISPs don't do anything about this, but it would make me feel better knowing that I have told someone, potentially (via their ISP), that their machine is infected. Many ISPs have an abuse email address that comes up on my whois query results.
PO 12/5/2012 | 2:04:54 AM
re: Mobile Skype: Quality Issues? A VoIP call can interconnect to the PSTN through a gateway. Whether this gateway is owned by a carrier or not, it is a line termination (phone), trunk termination (PBX) or a switch termination. So the PSTN network will use the standard protection mechanisms that are in place to handle overload. If the concern is that with VoIP, calls can be generated without a human intervention, isn't the situation similar to have a virus program in a computer that is designed to generate modem calls?

A few comments. If an individual line goes on/off hook enough over a particular time interval, the Class 5 should start to ignore that line and run diagnostics to see what's wrong. But if a large number of lines originate calls simultaneously then the CPUs in the Class 5 will start to deny originations. This is because these things were designed when memory & horsepower were expensive, so there simply isn't enough capacity to handle everything happening at once. Finally, of course, if the trunks are all occupied then again new originations will fail.

With VoIP, there is no longer a need for a person to "go off-hook" in order to generate an event aimed at the PSTN (via the PSTN gateway). Yes, this is an exposure; the gateway capacity can be subject to a DoS attack. And depending on the nature of an attack, parts of the PSTN beyond that gateway could also be cowed.

I don't know what steps operators such as Net2Phone have taken to protect themselves, but I would hope that they're relying on more than just not getting attacked by the script kiddies.
PO 12/5/2012 | 2:04:54 AM
re: Mobile Skype: Quality Issues? Doesn't seem like the network will be a mere collection of stupid pipes, does it?

Depends on what part of whose network we're referring to, and at what layer.

There isn't such a thing today as "The Network": there's MCI's network(s), AOL's network(s), AT&T's network(s), etc. within AT&T there are regional and inter-regional networks, monitoring and control networks, and VPNs for corporate customers.

Then, of course, at the physical layer there are different distinctions.
PO 12/5/2012 | 2:04:55 AM
re: Mobile Skype: Quality Issues? My firewall also has a feature where I can do a whois on the offending IP address.

You'll want to get a copy of logresolve, which will do a reverse lookup from the IP addresses in a whole log file. It's part of the Apache webserver distribution (although I wouldn't run Apache on a Win98 machine), but you can also find it on its own (google logresolve). Of course, if your logs are in a different format than it expects you may need to do some massaging.
technonerd 12/5/2012 | 2:04:55 AM
re: Mobile Skype: Quality Issues? The fact that this level of maturity hasn't yet arrived suggests that the interfaces to the VoIP call controllers will be kept as internal interfaces to the carriers for the foreseeable future.

Interesting discussion. Doesn't seem like the network will be a mere collection of stupid pipes, does it? Perhaps Mr. Isenberg was a bit off base? Nah.
PO 12/5/2012 | 2:04:56 AM
re: Mobile Skype: Quality Issues? SIP does not maintain state for calls though the proxy. No memory is consumed and so it is not possible for the SIP proxy to run out of resources in trying to service multiple call attempts.

Let's start here. Memory isn't the only resource of interest. It would be easier to launch a DoS attack based on bandwidth consumption, squeezing out true traffic. That is, the zombies simply fire off max-sized packets at the open port. The fact they get thrown away quickly doesn't matter much; the service is still starved. And if the packets get thrown away earlier (i.e. by routers, based on a size rule), then the zombies are adjusted for the next attack.

The longer term solution includes use of cryptographic identity (e.g. SSL) and packet filters in front of the service node (e.g. in routers) doing some pre-examination of traffic (disallow any non-SSL traffic). Protection is not provided by the complexity of the algorithm but rather by the traceability of the authenticated identity.

The fact that this level of maturity hasn't yet arrived suggests that the interfaces to the VoIP call controllers will be kept as internal interfaces to the carriers for the foreseeable future.
technonerd 12/5/2012 | 2:04:58 AM
re: Mobile Skype: Quality Issues? WiFi is a nomadic service that will be of great use in specifc locations. That is why it is so popular within businesses.
It's not popular within businesses. It has a measure of popular with the IT geeks. Not with anyone else.


You give a very good explantaion of why 3G with its emphasis on purely mobile services with handoffs is so expensive. Broadband handoffs in moving vehicles is also of little interest to customers. Who really wants to make broadband connections when driving in a car.
This thread began with a discussion of using Wi-Fi for VoIP service. Unless you can do handoffs, it's foolish to imagine that VoIP over Wi-Fi is going to be anything other than a geek toy. Thus, your claim that Wi-Fi does more is false. It does less.

That, in and of itself, is not a show-stopper, but the lessness of Wi-Fi is such that anyone thinking of using it for VoIP in any serious way must be on drugs.


Hence the great interest in the cheap nomadic service created by WiFi/802.11. It is no more expensive that the provision of payphones that used to be ubiquitpus in all stores and restaurants before cell phones.
Wi-Fi service isn't particularly cheap. The carriers I've seen clip people for $20 or $30 a month. This is essentially a toy for geeks who want to surf the net on their laptop at Starbucks. Not exactly the stuff of the mass market.

I agree that, as an individual paying my own bill, if I were inclined to surf the net somewhere other than home I'd probably want to use Wi-Fi rather than 3G on the grounds of cost. But I'm just a technonerd not a technogeek, and certainly not a dreaded propellerhead, so I'm not interested in either one of these things.

The closest I'll ever get to using 802.11 will be inside the house. I'd think about using it to link one of those tablet computers to my cable modem. However, there are some problems:

1) When I went to Fry's and asked about tablet computers, they said they didn't carry them because the things get really hot and burn out. Typical of anything Microsoft: The shit never works the first time around. It barely works the second time around. It's minimally acceptable the third time around. Just when it's really stable they come out with an "upgrade" that doesn't work.

2) Networking in Windows XP is a nightmare if you're not a technogeek or propellerhead. So even if the 802.11 thingy will broadcast through my walls reliably, I'm not confident that I could get it to talk to my cable modem. And if I can't do that, do you really think I'd spend 1,200 bucks on a tablet computer that heats up and burns out? Nah.
aswath 12/5/2012 | 2:05:00 AM
re: Mobile Skype: Quality Issues? once those 'outgoing trunks' are used up, what happens then? Can people still make 911 calls?

911 calls use "CAMA trunks", which are seperate from public trunks. Also keep in mind that in PSTN switches are king. They can cut off any call at any time. They have rules to handle focused overloads and the like. So the mantra is: There are no additional operational concerns introduced because of interworking with VoIP.
stephenpcooke 12/5/2012 | 2:05:01 AM
re: Mobile Skype: Quality Issues? Not possible. The Class 5 and the toll switches will accept only those calls for which outgoing trunks are available and the switches are designed to handle that much volume.

Forgive my ignorance once again, but once those 'outgoing trunks' are used up, what happens then? Can people still make 911 calls?
aswath 12/5/2012 | 2:05:02 AM
re: Mobile Skype: Quality Issues? if it was my goal to bring down a bunch of toll switches across the US I'd have all the domestic gateways max'ed out on every class 5 making long distance calls to POTS lines, from every country with a gateway. Possible?

Not possible. The Class 5 and the toll switches will accept only those calls for which outgoing trunks are available and the switches are designed to handle that much volume. What is possible is to program the gateways to go offhook and onhook many times in quick succession, simulating stuttering line (analogous to DoS attack). Class 5 switch protects itself in a procedural way, by putting that line out of service. In summary, PSTN has enough safeguard in place against mesbehaving VoIP terminals.
stephenpcooke 12/5/2012 | 2:05:02 AM
re: Mobile Skype: Quality Issues? Not possible. The Class 5 and the toll switches will accept only those calls for which outgoing trunks are available and the switches are designed to handle that much volume.

Forgive my ignorance once again, but once those 'outgoing trunks' are used up, what happens then? Can people still make 911 calls?
stephenpcooke 12/5/2012 | 2:05:03 AM
re: Mobile Skype: Quality Issues? dljvjbsl, perhaps you can describe this limitation? I'm no genius but if it was my goal to bring down a bunch of toll switches across the US I'd have all the domestic gateways max'ed out on every class 5 making long distance calls to POTS lines, from every country with a gateway.

Possible?
dljvjbsl 12/5/2012 | 2:05:04 AM
re: Mobile Skype: Quality Issues?
To my mind

... such as it is


You really cannot help yourself can you.

WiFi is a nomadic service that will be of great use in specifc locations. That is why it is so popular within businesses.

You give a very good explantaion of why 3G with its emphasis on purely mobile services with handoffs is so expensive. Broadband handoffs in moving vehicles is also of little interest to customers. Who really wants to make broadband connections when driving in a car.

Hence the great interest in the cheap nomadic service created by WiFi/802.11. It is no more expensive that the provision of payphones that used to be ubiquitpus in all stores and restaurants before cell phones.
dljvjbsl 12/5/2012 | 2:05:04 AM
re: Mobile Skype: Quality Issues?
Aren't the gateways the physical barrier?


The number of trunks between the gateway and the PSTN is limited. This would seem to limit the effects of any mass calling event from a gateway.
stephenpcooke 12/5/2012 | 2:05:05 AM
re: Mobile Skype: Quality Issues? I have no idea. That is why I am asking the question. Given the capability of Internet-based operations (ie: many simultaneous processes ongoing on a single computer, and many, many computers), the connections are virtual and do not tie up resources on the gateways (from previous posts). I would also guess that the capacity of the combined gateways surrounding a class 5 switch is greater than the capacity of that switch in terms of circuits or potential circuits.

Your point about multiple infected computers dialing modems is much the same but, I would submit, much more difficult to control. However, the common denominator would be immediately past the modem in the Internet connection itself (ie: once connected to the Internet in any way you could affect more circuits than the just the one used by the phone line and modem).
technonerd 12/5/2012 | 2:05:06 AM
re: Mobile Skype: Quality Issues? With an IP service you can have a single connection from multiple 'zombie' computers that can each generate multiple call requests, no real 'physical barrier' exists anymore.

I read the original post and all the replies in this thread, and rather than replying one by one I thought I'd jump in here. Stephen, you raise an outstanding point, and one that there is currently no answer for.

All you can really do is run your VoIP service through a computer hooked up to a good firewall, using an ISP that aggressively screens out spam. Even then, some of it will get through. Perhaps if VoIP ever becomes real in the consumer space, there'll be an expansion of the federal "do not call" list to VoIP.

In the meantime, anyone using VoIP will be a propellerhead science project. Which is one more of the many reasons to be a late adopter of this particular, um, "service."
technonerd 12/5/2012 | 2:05:06 AM
re: Mobile Skype: Quality Issues? You gotta be joking. Blackberry is the road warrior tool of choice. I hate the stinkin' things but they're being used by executives who make Dilbert's Boss look like a rocket scientist. Just what I need... inane email from some senior VP at 2AM.

That's funny, and trust me I do know what you mean. I remember getting a Blackberry. I eagerly hooked it up, thinking wow is this cool. Three weeks later I disconnected it because the f'ing thing drove me crazy.

Whenever I left the office, I put an autoreply on my e-mail saying that I had disconnected my Blackberry. I would check my e-mail once a day from my hotel, and if anyone needed to reach me urgently they should just call my office because it would forward to my cellphone regardless of where in the world I was.

It worked beautifully. Once I retired and they wanted me to return my Blackberry I went searching for it and found it on a shelf at home with a coating of dust. I wrapped it carefully, dust and all, and sent it back to the IT people. When they got it, they called me and said they wished there had been more people like me in the company.

More seriously, yes I know that Blackberries are used by a cohort of anal-retentive business people. They are real productivity destroyers. But in any case, by comparison to the number of cellular subscribers it's a tiny number of people who are using these things. Virtually no one pays for a Blackberry themselves. It's always a company-provided device.
technonerd 12/5/2012 | 2:05:07 AM
re: Mobile Skype: Quality Issues? However WiFi can accomplish much more than voice and much more than the cellular network is capable of.
This is the usual chatter of the propellerheads and propellerhead wannabes. It is based on very shallow thinking about "Wi-Fi" that's very reminiscent of equally shallow thinking about "DSL" that led to the loss of tens of billions of dollars during the Bubble frauds.

Like DSL, Wi-Fi is a marketing term. Wi-Fi has come to stand for anything operating over one of several 802.11 wireless standards. The key attributes of these standards are:

1) Unlicensed frequency
2) Broadband capable
3) Very high frequencies
4) No cell handoffs

At the moment, posters here focus on #2 but what matters most are #1 and #3 and #4. Wi-Fi being unlicensed means that its ability to be used for a high-volume commercial service is limited. In the incredibly unlikely event that a Wi-Fi carrier ever really did damage to a cellular carrier, the cellular carrier could force it to sell out by threatening to interfere. Nothing in the law could stop it.

But #3 is by far the most crucial weakness. Wi-Fi works on microwave frequencies with limited propogation. To compete effectively with cellular you'd need comprehensive coverage, and to do that with Wi-Fi you'd need many millions of cells. Each of these cells would have to be a base station connected to the Internet. It would be incredibly expensive for any carrier to blanket the country with Wi-Fi.

Problem #4 is also a voice killer. Wi-Fi doesn't allow for handoffs, which means it can only be portable rather than mobile. What does someone do with their Wi-Fi phone in their car? Nothing. It won't work in their car, because the car is moving and you can't do handoffs. Not only that, but if they ever did develop handoff capability could you just imagine the difficulty of managing the number of handoffs involved in a typical mobile session given the small sizes of cells?

Why am I reminded of DSL here? Because when CLECs were installing DSL (or claiming to), neither they nor the propellerheads in places like [i]Light Reading[/i] and the other trade journals ever bothered to look at the details. If they had done so, they'd have used different line codes. If outsiders had ever bothered to look, they'd have realized that the CLECs were making a fatal mistake when they used 2B1Q instead of QAM.

But that level of diligence required, well, diligence. And for VC-backed trade journals it required a level of integrity that they did not have, do not have, and will not have. If some VCs ever climb on the 802.11 voice bandwagon in a serious way and some IPOs loom, I have no doubt that Wall Street and the trade rags will do exactly what they did in the late '90s with respect to DSL: Spread any lie they can get away with, regardless of how outrageous.


To my mind
... such as it is ...


the failure of the 3G network is not purely technological but is instead more of a reflection of the inability of telecom-type networks to supply services fitted to the needs of particular users. 3G does what it does but it is too little compared to what can be done with other technology.
More ignorance talking here. 3G is another marketing term. It refers to wideband services delivered on sub-microwave high frequencies. The technology works fine, but it's expensive. Especially if you put it on the upper band of cellular, which turns out to be the lower end of microwave but not quite microwave, i.e., what was once called the "PCS" frequencies.

Put 3G there, and you need four times as many cells to get the same coverage as you do with the 800 HMz frequencies. But you pretty much have to use the higher frequencies for 3G, because you need all the spectrum you can get. Which means a lot of cells. Very densely packed areas can support 3G because they have lots of cell sites anyway, i.e., South Korea and Japan.

Less dense areas, much of Europe and all of the United States, have a hard time with 3G because of this issue. It's not the technology, it's the cost. You need more cellsites for 3G -- a lot more -- and that means you require paying customers for those services. A lot of them. There just aren't enough people in the U.S. who care about mobile or even portable wideband and broadband services to pay for the expansion of the 3G-capable network.

The Wi-Fi phemonenon, i.e., using microwave frequencies instead of sub-microwave frequencies for wideband and broadband packet service, is a tiny phenomenon of appeal almost exclusively to a handful of geeks. To the extent that, say, traveling businessmen might want to use such a service, the cellular carriers are going into the airports and similar locations where those laptops can be served.

And frankly, using the 3G cellular network (when available) is a lot more convenient than using Wi-Fi. Not because of any technology difference, but because logon and registration procedures are a whole lot easier. This begs the question of why anyone bothers with public Wi-Fi, and that leads back to it being something by propellerhead for propellerheads. Public Wi-Fi was funded right at the peak of the bubble by a handful of VCs in the blindest of their blindness.

Public Wi-Fi loses money, and will always lose money. The only question is when the plug will be pulled. Propellerheads who keep throwing money at Wi-Fi services really ought to ask themselves if there aren't some more deserving charities out there.
aswath 12/5/2012 | 2:05:07 AM
re: Mobile Skype: Quality Issues? Aren't the gateways the physical barrier?

For modem calls, I was thinking of multiple infected computers, along the lines of DDoS.
stephenpcooke 12/5/2012 | 2:05:08 AM
re: Mobile Skype: Quality Issues? Aswath, your understanding and mine are somewhat similar in this except that, to generate modem calls you have to have multiple physical phone lines and physical modems connected to them to overload a switch. With an IP service you can have a single connection from multiple 'zombie' computers that can each generate multiple call requests, no real 'physical barrier' exists anymore.
aswath 12/5/2012 | 2:05:10 AM
re: Mobile Skype: Quality Issues? stephenpcooke:

I do not see the problem for the following reasons:

A VoIP call can interconnect to the PSTN through a gateway. Whether this gateway is owned by a carrier or not, it is a line termination (phone), trunk termination (PBX) or a switch termination. So the PSTN network will use the standard protection mechanisms that are in place to handle overload. If the concern is that with VoIP, calls can be generated without a human intervention, isn't the situation similar to have a virus program in a computer that is designed to generate modem calls?

What am I missing?

Aswath
stephenpcooke 12/5/2012 | 2:05:12 AM
re: Mobile Skype: Quality Issues? SIP does not maintain state for calls though the proxy. No memory is consumed and so it is not possible for the SIP proxy to run out of resources in trying to service multiple call attempts.

How does this relate to class 5 circuits?
stephenpcooke 12/5/2012 | 2:05:12 AM
re: Mobile Skype: Quality Issues? dljvjbsl wrote:

I should ahve mentioned that the availablity requirements for telpehone networks are much more subtle thanhte 5 9s that is always mentioned. However even with that there are vunerabiliites in the network for major outages. The SS7 overload message bug which shut down the network in the eastern US several years ago is an example of this. I know of no similar occurence in the Internet.

Even DDOS attacks target only single locaions and can be defeated by modifying the routing in the core. Single bugs do not seem to be able to shut down the network as has happened in the PSTN


I may be one of the few who understands the availability requirements in NEBS better than most. The test company that I worked for was a NEBs house and was Verizon certified. I was also the troubleshooter of last resort on Nortel's TransportNode product line and have some experience with network outages. The troubling part about your comment for me is that the outage that you describe was not a 'known and widely published' difficulty prior to the event. Yes, there are vulnerabilities in the PSTN but they are extremely difficult to find and it is even more difficult to make purposeful use of them on a large scale. Just because you know of no DDOS attacks that have gone after more than one location does not mean that there haven't been any or cannot be any in the future. SIP interfaces to class 5's are not 'generally unknown or unpublished' PSTN vulnerabilities, these are widely published standards available to everyone with Internet access. Also, someone who REALLY wanted to could quite conceivably go after many different bugs all at once, just because they are there and published freely. Of course it helps that they can write their own interface to a class 5 and distribute it via the Internet incredibly easily and they can have a whole bunch of computers do the damage for them.
dljvjbsl 12/5/2012 | 2:05:13 AM
re: Mobile Skype: Quality Issues?
'next to no resource on the proxy'


SIP does not maintain state for calls though the proxy. No memory is consumed and so it is not possible for the SIP proxy to run out of resources in trying to service multiple call attempts.

Trust realtionships can also be set up so that the proxy will ignore messages from unknow sources.

This issue has been adressed by the 'in-crowd' who control the SIP standards.
dljvjbsl 12/5/2012 | 2:05:15 AM
re: Mobile Skype: Quality Issues? I should ahve mentioned that the availablity requirements for telpehone networks are much more subtle thanhte 5 9s that is always mentioned. However even with that there are vunerabiliites in the network for major outages. The SS7 overload message bug which shut down the network in the eastern US several years ago is an example of this. I know of no similar occurence in the Internet.

Even DDOS attacks target only single locaions and can be defeated by modifying the routing in the core. Single bugs do not seem to be able to shut down the network as has happened in the PSTN
stephenpcooke 12/5/2012 | 2:05:15 AM
re: Mobile Skype: Quality Issues? Also, at what point does a circuit get set up on the class 5? How many messages does this require and what is stopping them from overloading that class 5? Thanks.
stephenpcooke 12/5/2012 | 2:05:17 AM
re: Mobile Skype: Quality Issues? 'next to no resource on the proxy' what does this mean exactly and how does it work when applied to thousands of computers all sending these messages?
dljvjbsl 12/5/2012 | 2:05:18 AM
re: Mobile Skype: Quality Issues?
What is stopping some organized hackers from creating a network of 'zombie' computers (via worms, adware, spyware, etc.) to perform DDoS or similar attacks on our phone system?


SIP has been designed as a stateless protocol with this contingency in mind. A SIP message consumes next to no resource on the proxy and so it is next to impossible to send enough messages to hinder its operation.
stephenpcooke 12/5/2012 | 2:05:18 AM
re: Mobile Skype: Quality Issues? optical_man, thanks for the response. I guess the difficulty that I am having is that, as long as the POTS network is linked to any other network whose members can be programmed in any way and are accessible to the general public, the telephone network is vulnerable. I don't have faith in anyone being able to 'fix' this with another box and/or accompanying software.

The point is: no telecom network equipment that is connected to the Internet, in any way, can be classified as meeting the NEBS requirement of having 99.999% uptime in a year.
optical_man 12/5/2012 | 2:05:20 AM
re: Mobile Skype: Quality Issues? What is stopping some organized hackers from creating a network of 'zombie' computers (via worms, adware, spyware, etc.) to perform DDoS or similar attacks on our phone system?


SC,
Not much is stopping this right now. Some pretty bright minds see this as a business opp and are creating apps/hardware to secure voip networks. they will make some nice $$.
I think (however economically disruptive it is) it will be a hilarious situation when all the nations phones (or at least the max # of provisioned circuits within each switch) start ringing ALL AT ONCE!!
Funny for maybe 2 minutes, then we'll realize this is a huge wake up call, and the developers of voip security apps/hardware will FINALLY see PO's pour in by the thousands.
stephenpcooke 12/5/2012 | 2:05:20 AM
re: Mobile Skype: Quality Issues? Hi,

I have a firewall installed on my home Internet connection. I occasionally look at the logs. The firewall that I have is installed on an older Windows 98 machine (recycling, awesome!!!) so it has lots of disk space to store previous log files. I had a look at a few of those log files. On average, there were over 1000 illegal login attempts on my home network in every approximately 2-3 hours. That averages out to about one illegal login attempt every 6 seconds!!! My firewall also has a feature where I can do a whois on the offending IP address. This basically finds the ISP on whose network the offending machine resides and in some cases provides an 'abuse' contact email, but almost always gives me the country of the offending machine. Note that this does not count 'Adware' or 'spyware' programs that are unwittingly downloaded.

I had illegal login attempts from the Netherlands, Florida, Montevideo, Columbia, Australia, China, and (mostly) from within my own ISP's network. I only checked a few of the addresses.

So, my question goes like this to all VoIP advocates out there:

Given that:
- the approach is to make a dumb network (so that the network has little or no intrinsic intelligence to police the bits travelling through it)
- any VoIP network uses applications on the edge of the network (ie: resident within my computer or VoIP phone, etc.)
- most of those illegal connection attempts will have come from 'zombie' computers infected with Internet worms
- any VoIP network is not isolated because the interface to it is on the public Internet through an ISP
- there exists many ways in which I can make a phone call to someone on the POTS network through my computer over the public Internet (ie: IP to Class 5 gateways exist all over the place)

What is stopping some organized hackers from creating a network of 'zombie' computers (via worms, adware, spyware, etc.) to perform DDoS or similar attacks on our phone system?

Please explain.
dogmeat 12/5/2012 | 2:05:21 AM
re: Mobile Skype: Quality Issues? VoIP capable AP infrastructure.

Take a look a Meru Networks. They are the only ones in this space that appear to have solved the RF problems in 802.11.
alchemy 12/5/2012 | 2:05:23 AM
re: Mobile Skype: Quality Issues? dreamer101 writes:
1) Skype use Voice codec from Global IP sound that supposed to be able to tolerate 30% packet loss with no loss in voice quality. I do nto know whether it is true

iLBC is a very good compression codec but the voice quality does drop off significantly if you have sustained packet loss. For those who aren't familiar, the iLBC codec from Global IP Sound supports basic frame sizes of 20 and 30 mSec. It's toll quality at relatively low complexity compared to, say, G.728. It's quite good at dealing with dropped packets.

Given that Wifi, DSL, Cable Modem has a lot more bandwidths than it is required for voice. Waste bandwidth so it can works in a low quality transport situation looks more reasonable that required QOS in all networks. Bandwidth is cheap in those scenerio

For voice over cable service offered by the MSOs, they're going to eventually phase in compression codecs. Bandwidth is cheap until you do the analysis on running multiple simultaneous G.711 VoIP streams on a single DOCSIS upstream frequency. DOCSIS has big pipes on the downstream and small pipes on the upstream. DOCSIS 2.0 repairs this somewhat but there's no DOCSIS 2.0 plant deployed yet and cable modems are mostly at DOCSIS 1.0 and 1.1.

The 'bandwidth is cheap' argument works great on a PowerPoint slide deck but when you do the real traffic analysis on the access network, VoIP ends up being a real hog if you have 30% market penetration on homes passed and are traffic engineering using traditional telco erlang numbers. You end up having to double your DOCSIS cable modem plant if you assume a 10% take rate on cable modem and 30% take rate on telephony. 2x the number of fiber nodes. Twice the fiber. 2x the number of CMTSs. Truck rolls to partition the network and move amplifiers around.
alchemy 12/5/2012 | 2:05:23 AM
re: Mobile Skype: Quality Issues? technonerd writes:
By the way, RIMM and Treo are gearhead toys. Very few people use them.

You gotta be joking. Blackberry is the road warrior tool of choice. I hate the stinkin' things but they're being used by executives who make Dilbert's Boss look like a rocket scientist. Just what I need... inane email from some senior VP at 2AM.
dljvjbsl 12/5/2012 | 2:05:23 AM
re: Mobile Skype: Quality Issues?
The problem with WiFi (IMHO) remains in the business model: to date, nobody has figured out how to be successful deploying a commercial WiFi network of any breadth


It si now being offered as a complimentary service in hotels jsut as the hotels offer cable television and local phone service. These services were charged for at the beginning as well. I would expect that WiFi will be seen as a necessary amenity like street furniture in business areas and supplied accordingly
PO 12/5/2012 | 2:05:25 AM
re: Mobile Skype: Quality Issues? However WiFi can accomplish much more than voice and much more than the cellular network is capable of.

I see WiFi and cellular developing as complementary technologies. Lazaridis and friends at Research in Motion were quick to add telephony to their device. I don't doubt they're also working to add 802.11 capabilities.

The problem with WiFi (IMHO) remains in the business model: to date, nobody has figured out how to be successful deploying a commercial WiFi network of any breadth. Sure, they've been able to offer it as a sweetener to other sales, but until someone can make money deploying it, service levels will remain random at best.
sgan201 12/5/2012 | 2:05:25 AM
re: Mobile Skype: Quality Issues? Hi,
1) Skype use Voice codec from Global IP sound that supposed to be able to tolerate 30% packet loss with no loss in voice quality. I do nto know whether it is true. But, the approach is correct.. Given that Wifi, DSL, Cable Modem has a lot more bandwidths than it is required for voice. Waste bandwidth so it can works in a low quality transport situation looks more reasonable that required QOS in all networks. Bandwidth is cheap in those scenerio

2) USA is not a growth market for telecom. So, who cares whether VoWiFi applies in USA or not. All major advances in application in wireless happen outside of USA anyhow..

Dreamer
dljvjbsl 12/5/2012 | 2:05:27 AM
re: Mobile Skype: Quality Issues?
...
Cellular service costs $20-$30 a month and includes free LD. Wi-Fi cannot and will not compete against this
...


However WiFi can accomplish much more than voice and much more than the cellular network is capable of.

To my mind, the failure of the 3G network is not purely technological but is instead more of a reflection of the inability of telecom-type networks to supply services fitted to the needs of particular users. 3G does what it does but it is too little compared to what can be done with other technology.
technonerd 12/5/2012 | 2:05:28 AM
re: Mobile Skype: Quality Issues? Wifi voice would be killer of cellular industry in 5-10yrs time frame. Watch out. With all the popularity of devices like RIM & TREO, u will be seeing free voice calling using the portable wireless devices.

Future bodes a very murky picture for cellular providers with Wifi unless they adapt to a hybrid model with wifi used when in static locations and cellular when one is on the move.


There will be no "free" calls over WiFi other than a few from propellerheads who are playing around. In any case, even they will have to pay a subscription fee for Wi-Fi service, and their Wi-Fi provider will have to pay access charges on any calls routed through the PSTN. Want to call home on Wi-Fi? Ka-ching!

By the way, RIMM and Treo are gearhead toys. Very few people use them. Cellular service costs $20-$30 a month and includes free LD. Wi-Fi cannot and will not compete against this; it's absurd to imagine otherwise. But hey, there's always some stupid VC in California who you might convince. Give it a shot!
jcbogle 12/5/2012 | 2:05:28 AM
re: Mobile Skype: Quality Issues? Wifi voice would be killer of cellular industry in 5-10yrs time frame. Watch out. With all the popularity of devices like RIM & TREO, u will be seeing free voice calling using the portable wireless devices.

Future bodes a very murky picture for cellular providers with Wifi unless they adapt to a hybrid model with wifi used when in static locations and cellular when one is on the move.

---
The only way Wi-Fi voice could ever develop a user base other than a few propellerhead hobbyists would be if cellular systems became too crowded or expensive. And that's not likely to happen; in fact, all signs are pointing in the opposite direction
truelight 12/5/2012 | 2:05:33 AM
re: Mobile Skype: Quality Issues? CISCO does
Betelgeuse 12/5/2012 | 2:05:37 AM
re: Mobile Skype: Quality Issues? IG«÷m just starting to look into QoS enabled 802.11 for my company. Does any one currently offer a Wi-Fi switch that is QoS enabled? From what IG«÷ve read of 802.11e it seems to leverage CSMA/CD with some modifications. I donG«÷t see how a CSMA/CD system will ever be able to support latency and jitter sensitive IP applications. Are there any other plans in the works?
technonerd 12/5/2012 | 2:05:38 AM
re: Mobile Skype: Quality Issues? Voice over WiFi is a difficult problem and one that can not be solved at an application level. The basic problem is that the typical WiFi beacon rate is 20msec. So your best case is a 20 msec packet jitter

The only way Wi-Fi voice could ever develop a user base other than a few propellerhead hobbyists would be if cellular systems became too crowded or expensive. And that's not likely to happen; in fact, all signs are pointing in the opposite direction.
optoslob 12/5/2012 | 2:05:38 AM
re: Mobile Skype: Quality Issues? Voice over WiFi is a difficult problem and one that can not be solved at an application level. The basic problem is that the typical WiFi beacon rate is 20msec. So your best case is a 20 msec packet jitter, this is almost tolerable but the jitter problem gets worse as the node gets loaded. Most people that are doing VoWiFi are modifying the algorithms in the access point to give preferential treatment to the voice applications. Even if this is done you still have a hand off problem and other problems with microwave ovens and BT interference. Most of these problems will be solved with a/b/g systems and useable systems will be possible with n/e.

optoslob

materialgirl 12/5/2012 | 2:05:40 AM
re: Mobile Skype: Quality Issues? This is just the start of the big wireline bypass. As Clayton Christensen says, disruptive technologies always look pathetic and spotty at first, used by fringe groups. Your arguements against this one are typical, and irrelevant.

The fact is that Skype has the economic model to grow. And grow it will. Be patient. Look back in a year, and you will be startled.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE