x
Services

QOS Fees Could Change Everything

The revelation that BellSouth Corp. (NYSE: BLS) is looking to charge content providers – especially gaming companies and movie download sites – for a premium ride across its network is good news for several gear vendors. But, at the same time, the trend is causing a stir among standalone VOIP companies.

BellSouth's plans, first reported last week in the Wall Street Journal, hits the topic of "net neutrality" – a belief that owners of broadband networks shouldn't discriminate against the kinds of applications and content sent across their network. (See Chirac Loves FTTH and Brand X Decision Stokes VOIP Worries.)

The notion that service providers would charge extra for quality of service (QOS) challenges the idea of net neutrality, and could put third-parties using broadband connections for VOIP and other services demanding low-latency connection at a distinct disadvantage.

"Broadband providers are taking an unpopular stance and saying, 'We won't provide QOS for any other services but ours,' " says Benoit Legault, VP of marketing at Ellacoya Networks.

The move, of course, could provide a boost to those producing "packet inspection devices," or traffic management. Deep packet inspection technologies, such as those sold by Caspian Networks Inc. , Cisco Systems Inc. (Nasdaq: CSCO), Ellacoya Networks Inc. , and Juniper Networks Inc. (NYSE: JNPR), to name a few, would be in high demand once a service provider decides to give paying content providers traffic priority over emails, file downloads, and even consumer VOIP phone calls.

VOIP providers, meanwhile, aren't so thrilled.

Critics say the phone companies are cooking up these new fees as a way to book more revenues without significantly upgrading their networks. Bryan R. Martin, CEO of Packet 8, is excited that BellSouth is reselling his company's VOIP service, but that doesn’t mean he agrees with the idea of carrier QOS fees. "We're hopeful that the networks will get better, but if one of these companies were to approach 8X8 and say, 'Would you pay X many dollars a month in order to get some sort of improved access' – be it bandwidth or latency or whatever – we would just laugh at them and say, 'No,' " says Martin.

Proponents of net neutrality worry that if BellSouth starts charging for premium access, it may decide to limit what kinds of content users have access to – or it may charge users more money for sharing home videos or other high-bandwidth applications.

BellSouth spokesman Jeff Battcher says that his company isn't interested in being the content police or trying to limit what people can do on the Net. Rather, he says, they're trying to provide a service in the video world that's analogous to what companies get when they buy a 1-800 telephone number.

"When you call a 1-800 number to Lands End… they're willing and want to pay for that call so they can receive your business," Battcher says. He says that similarly, companies are approaching BellSouth to ask for a faster pipe to some consumers so those consumers won't be disappointed and, in turn, blame the content provider for a bad experience.

But if content providers don't pay for BellSouth's premium service, then nothing much will change, Battcher says.

BellSouth and other carriers haven't arrived at a firm conclusion as to whom they'll charge, whom they won't charge, and what kind of money is at stake.

BellSouth's Battcher says that none of BellSouth's deals with gaming companies and content providers will involve making the consumer pay more for a DSL connection. However, the very technology that BellSouth would use to deliver on those deals can be applied in reverse to allow consumers to pay for brief periods of higher bandwidth bursts. (See BellSouth's Smith Details IPTV Plans.)

Deep packet inspection companies say some network congestion problems can be solve by broadband providers looking at their network's traffic and applying some "proportional enforcement" to those heavy users of, for instance, P2P file-swapping services.

Junaid Islam, VP of marketing at Caspian, says in China one big carrier used proportional enforcement to keep P2P traffic from ruining the experience of casual Web users. The more often users downloaded large files during peak usage hours, the slower the downloads would occur. No traffic was blocked, but the slowdowns forced businesses and other heavy downloaders to either upgrade their services or wait to do their heavy use until odd hours.

In the U.S., deep packet inspection companies say that some very large carriers are looking at applying policy-based routing and other ways of building in QOS throughout their networks. And, to the chagrin of VOIP companies, the early discussions suggest that such services won't be free.

One motivating factor in going to the QOS model is that it would come before broadband has really hit its growth peak – and thus it would ensure better profitability going forward. For example, a carrier like BellSouth has hooked up less than 20 percent of the potential customers it can reach.

BellSouth last reported that it had 2.7 million DSL customers, and the carrier will update those figures during its next earnings call on January 25. Some 85 percent of the 16 million households the carrier serves – about 13.6 million households – can be reached by at least a basic 1.5-Mbit/s DSL connection.

How do you feel about this issue? Take our latest poll: Net Neutrality.

— Phil Harvey, News Editor, Light Reading

cloudyskyhere 12/5/2012 | 4:07:06 AM
re: QOS Fees Could Change Everything I'm a hardware engineer. I'm very interested for the hw job there. Anyone can pass me a lead so I can have my resume turned in? :-D

You can use this email:[email protected]

Thanks, Cloudy
OpticOm 12/5/2012 | 4:08:03 AM
re: QOS Fees Could Change Everything I would appreciate any info regarding the HW hiring VP.
green 12/5/2012 | 4:08:05 AM
re: QOS Fees Could Change Everything I heard they are doing some storage virtualization stuff.

if anyone has the HR or SW eng hiring contact please post it. thx for the info.
eawest 12/5/2012 | 4:08:10 AM
re: QOS Fees Could Change Everything Is there a Web site for this new company or contact info?
Outsider 12/5/2012 | 4:08:11 AM
re: QOS Fees Could Change Everything Who in particular are you looking for?
OpticOm 12/5/2012 | 4:08:11 AM
re: QOS Fees Could Change Everything The hiring manager for the HW group.
OpticOm 12/5/2012 | 4:08:12 AM
re: QOS Fees Could Change Everything Does anyone have the contact info for the HW VP?
That would be very much appreciated.
Outsider 12/5/2012 | 4:08:12 AM
re: QOS Fees Could Change Everything Ron202,

Do you know this SW Director's name that joined Nuevo Technologies?

sgan201 12/5/2012 | 4:08:14 AM
re: QOS Fees Could Change Everything Network congestion

1) Moderate network congestion
Packet flowing into a box is at a higher rate than the output trunk port. The packet is being bufferred at the egress queue of the output port. The packets are delayed but not thrown

2) Severe network congestion

Pakcets are flowing into a box at a a higher rate than the output trunk port. The packet is coming at a rate where the egress queue will be full and packets has to be thrown to avoid overrrun. Normally, this is set at 80% point of the egress queue.. During severe network congestion, packet will be lost.

Dreamer
PO 12/5/2012 | 4:08:15 AM
re: QOS Fees Could Change Everything "... During network congestion ..."

How would you define this?
Outsider 12/5/2012 | 4:08:16 AM
re: QOS Fees Could Change Everything The name is Nueovo Technologies.

I heard today that they hired some people from the hardware team here at Cisco.

However I don't believe that they can hire any more people from Cisco.

Too bad..............

net_monk 12/5/2012 | 4:08:16 AM
re: QOS Fees Could Change Everything Is there a name for this company?
ron202 12/5/2012 | 4:08:16 AM
re: QOS Fees Could Change Everything last one I heard was a sw director...
ron202 12/5/2012 | 4:08:16 AM
re: QOS Fees Could Change Everything no , it is not true.
curiousgeorge 12/5/2012 | 4:08:17 AM
re: QOS Fees Could Change Everything Yes it is very real, and is into 'data center virualization'. They've added some credible server architects to their team, so it is not a bunch of networking folks only.

The person I know there cannot tell me more without either hiring me or killing me :-). Neither is an option at the moment, so...
Outsider 12/5/2012 | 4:08:17 AM
re: QOS Fees Could Change Everything I heard that they are not permitted to hire anyone from Cisco, is this correct?

If so I will quit my job at Cisco today........
Pete Baldwin 12/5/2012 | 4:08:17 AM
re: QOS Fees Could Change Everything Yes, we've been trying to peek into this one. It looks like they're working on the next Cisco spin-in, although we haven't sussed out exactly what they're doing yet. Could be anything, I suppose.

We've heard it's a self-funded venture so far -- Mario, Luca, et.al. have turned down all offers from VCs, or so we're told.

I guess this is their idea of retirement.
Outsider 12/5/2012 | 4:08:18 AM
re: QOS Fees Could Change Everything Is anyone aware of a new startup company in the Silicon Valley headed up by 4 Cisco Systems superstars- Prem Jain, Mario Mazzola, Soni Jiandani, Luca Cafiero?

A very reputable recruiter who I have known for years just recently contacted me regarding this company. Based on my own web search I thought that all of these guys retired from Cisco this past summer.

Any info will be appreciated.
Mark Seery 12/5/2012 | 4:08:21 AM
re: QOS Fees Could Change Everything Steve,

Important improvements in Ethernet efficiency / performance have included:

-moving from Manchester encoding to 4B/5B and 8B/10B (which you may be familiar with).

-moving from many devices per logical segment (hub or cable) to two(for example point to point)

-the addition of full-duplex operation (not present in early versions of Ethernet)

-increase in speed of course

-cut through switching (switching before entire frame received).

So it was not simply enough to go to point to point links, the addition of a full-duplex mode was vitally important to getting to the kind of performance you are used to seeing. Manchester encoding was very inefficient which did not matter so much on coax, but did on twisted pair (I have been told). As late as the mid-90's I was using Ethernet networks that were only getting 30% utilization (I know this from personal verification with analyzers) due to collision recovery and many users on the same segment; but though you might read about such things, most people consider this so far in the distant past, that your original comments did not resonate.

If you are using a full duplex Ethernet link to a customer's switch or router, you should see pretty good performance. Also note that the minimum frame size on Ethernet is 64 bytes, so 40 byte TCP acks (a common source of small packets) have to be slightly padded out (after adding IP header) so there is a small amount inefficiency there in addition to the previously mentioned interframe gap (which is very small at high speeds). This small amount of inefficiency may also impact the transfer of IP packets from a line card to a switching fabric when using a fixed size cell (for example 64 bytes). Some designs will exhibit corner cases around many small packets, and hence "hero tests" while unrealistic, test for this corner case (smallest possible packet size at wire speed on as many interfaces as possible). There are many different packet sizes on a typical network, so while TCP acks are not infrequent, they are not so frequent as to make a test consisting just of this packet type a reflection of real world conditions. Some people refer to something called a "IMIX" which comes in simple forms (a few packet sizes) or more complex forms with a greater variety in packet sizes.

Which is all just a long winded way of saying that many Ethernet switched and IP routed networks today exhibit low latency, low jitter, and high utilization (note some theory suggests that utlization over 70% will lead to queues becoming deeper - and hence capacity managers may plan to avoid this; compare and contrast this to a TDM or ATM VC network where capacity is unused because each individual channel is not fully utilized from a offered payload perspective by its client signal/traffic demand).
stephencooke 12/5/2012 | 4:08:21 AM
re: QOS Fees Could Change Everything Thanks for this Mark, and others. It is appreciated.

Steve.
sgan201 12/5/2012 | 4:08:22 AM
re: QOS Fees Could Change Everything PO,

1) I think you missed a very key point on my proposal. What I am suggesting is per user/subscriber level queueing and traffic management. During network congestion, each user/subscriber will be traffic shaped to their guarantee rate and each subscriber has its own queue/buffer. So, people substantially overused their 20Kbps will have their buffer overrun and lose a lot of packets. Meanwhile, for a person that only use 20Kbps, he/she will not lose any packets even under severe network congestion. This is what I called fairness from the customer standpoint. Why should a well behaved customer suffer from the abusive behavior of some customers.

2) The major difference between a SOHO user and a consumer user is that a SOHO user would like to have a guarantee levell of some bandwidth so that they can do their job/business at all time.

3) The amount of per user/subscriber/connection buffer required for this is not high at all. And, it had been implemented on some low cost boxes.

Dreamer

BigBrother 12/5/2012 | 4:08:23 AM
re: QOS Fees Could Change Everything There are many network procesors out there that can take in 10G traffic and split them into smaller pieces. That is why Intel is also getting into the NP game. Also a lot of the boxes that do packet inspection that uses software and FPGA combination, they can almost achieve full line rate in most cases except the extreme 64 bytes but then most of the flows are bigger than 64 bytes. Non-software solution is faster but there are their limit, changes are slow and more expensive.
OldPOTS 12/5/2012 | 4:08:24 AM
re: QOS Fees Could Change Everything Several routers and ethernet switches at the DSLAM/OLT are capable of operating at at least 10G. Most of these can do very deep packet inspection. A couple of years ago they were not yet at a price point ($xk) that operators would place as CPE on site. But these have now been produced in much larger volumes, and I assume have been cost reduced. But this is why operators must go after those $100/m customers that buy premium services.

BTW - These also have great capabilities to enforce SLAs, but you may want to place them where you have heavy traffic generators and let those 486s control/limit the rest. Those at the DSLAM/OLT can monitor/alarm when somone becomes a heavy traffic generator.

OldPOTS

PO 12/5/2012 | 4:08:24 AM
re: QOS Fees Could Change Everything "People are allocated 20Kbps. But, it is not guaranteed. How much would it costs to make those 20Kbps GUARANTEED as opposed to only ALLOCATED?? My contention is it does not costs that much."

Others have addressed this, but the question still comes back. Let me take another stab at this.

We need to consider each direction of traffic flow in a few different scenarios.

Let's start with outbound TCP traffic from a subscriber, across a DSL link and the Greater Internet, to a high-capacity site in a completely uncongested network.

TCP is rate adaptive, and wants to find the upper limit to the bandwidth available. But what does this really mean? It means that packets will queue up to traverse a slow link, and will zip along other (faster) links. In our scenario, the slowest link is the DSL uplink, so packets will queue up at the DSL modem until its buffer overflows, indicating (roughly) that the maximum available end-to-end bandwidth is being used for that flow.

If we start a second TCP flow across that same congested link, they'll generally each get half of the available bandwidth. (Assuming the TCP stack is correctly implemented; there are some which do not behave as well as others.)

Now if we start to add a moderate level of congestion at various points along the paths of those TCP flows, we'll start to add additional points where packets might be queued (delayed), though not yet dropped. So although the round-trip time (RTT) might increase slightly, the bandwidth for the TCP isn't decreased.

Then if we add more congestion, we'll start to drop packets at some other point(s) along the path, with the most heavily congested point weighing the most in the determination of how much bandwidth the TCP session will rate-adjust to.

Somewhere between those last two scenarios, the network starts to take proactive action, using tools such as Random Early Discard (RED) to prevent against something known as congestion collapse, or perhaps Explicit Congestion Notification (ECN). We won't worry about congestion collapse, though, beyond knowing that RED is a good thing.

Where is that congestion likely to occur? At a point in the network which the operator is economically incented to maintain at maximum occupancy. That is, the most expensive part of the network, which expense the operator doesn't want to waste. Typically, that'll be the Network Access Point (NAP).

But what if there's "excess capacity" in the rest of the network, so that all the access traffic arrives at the NAP only to have some of it dropped there due to congestion? The carrier has wasted money transporting those packets across the network, and has possibly forced other packets to drop which otherwise could have been delivered.

So it's in the carrier's economic interest, and in the interest of the traffic, for the access devices to be congested (concentrated) at a level comparable to the congestion at the NAP.

(A further grace is often available in that transport bandwidth is typically symmetric, and there is typically a greater demand for download bandwidth, as DSL and cable link capacities are much higher in the downstream direction.)

And for the transport network lying between the access and the NAP to not be overly congested either: losing packets in that section of the network wastes access and NAP resources.

Let's turn our attention to the return path, to traffic arriving at a subscriber's DSL link from some far-away high-capacity site. In an uncongested network, the "skinny pipe" is still the downlink of the DSL head-end, and packets queue for transmission. Eventually the buffer overflows and TCP rate-adapts to that effective bandwidth.

Again, we start to add moderate congestion across the network and find the packets-in-flight queuing at different points along the path, without significantly changing the Bandwidth-Delay Product.

Additional delay creates a new "choke-point" and the effective bandwidth for that TCP flow is reduced, with excess capacity on the DSL link available to other TCP flows.

If we start to add additional TCP flows, however, they will each adjust their bandwidth accordingly, and the bandwidth of the first flow will be reduced.

But again we note that it is in the economic interests of the carrier to maintain the overall level of congestion at the DSL head-end at a comparable level to the congestion at the NAP: paying for buffers that are never used is wasteful, as is paying for capacity that is never used.

What about traffic that doesn't behave well? That doesn't rate-adapt? Well, quite simply, it gets crappy service: it'll experience excessive packet loss, and high delay. And it makes life worse for everyone else, too: it squeezes out other traffic which otherwise could have been successfully delivered.

Now, what about that "guarantee" of, say, 20Kbps? To what purpose? If the traffic will still rate-adapt down to a fair-share allocation based on other congestion points in the network, how much good would the guarantee be? And if the traffic doesn't rate adapt, it'll still be lost at those other congestion points.

And it raises the networking conundrum: which is better? 100 TCP flows across a 20Kbps link, or 1 TCP flow across that same 20Kbps link? Different scenarios will treat them differently, because the Internet embeds a certain concept of "fairness" to traffic management, which percolates through TCP/IP (and UDP), and BGP routing, among other internet protocols.

Of course, networks typically have thousands of subscribers and multiple NAPs to choose from, but I would still stand by the concepts above for capacity management.

This doesn't say that QoS proposals are either a good thing or a bad thing: I've never made any such judgement. All it does is begin to outline the vast scope of issues which need to be addressed whenever the issue of QoS comes up.

I hope it helps.
PO 12/5/2012 | 4:08:24 AM
re: QOS Fees Could Change Everything "Bottom line is that comparing telecom bandwidths with actual IP data throughput numbers is not a fair fight. Unless vendors put beefier (ie: faster than line rate) processors on their boxes transmission efficiencies will not significantly improve from what they are now, as far as I can see."

This is a different question, of course. But it's still missing significant context.

Packets have various headers, and a "content" (protocol data unit, or PDU) section. It's generally only parts of the headers that processors have to look at, so the question becomes one of packet rate (or, inversely, packet size at line rate). On the internet, you'll see a few popular sizes around 64 bytes, around 512 bytes, and around 1500 bytes. The most stress for the processor is all 64-byte packets, at line rate.

For lower line rates (10, 100Mbps) you can get cost-effective processors to do the work. For higher line rates, or more involved inspection capabilities you would typically look to some sort of hardware assistance (FPGA, DSP, similar solutions).

There has historically been a tension in equipment design around this point: you can overbuild your equipment to handle 64-byte packets at line rate, then never see such a scenario outside the lab; or you can build to cost targets and underperform in the lab evaluation.

Most vendors today will find the additional costs acceptable and will build to line rate.
PO 12/5/2012 | 4:08:24 AM
re: QOS Fees Could Change Everything "So with today's IP boxes and LAN/WAN topologies, the effective line utilization gets MUCH higher than we were all taught was possible during the 10base-T days."

10Base-T was always capable of pretty close to line rate. (There's an inter-frame gap, but other than that you can go to town.) Even 10Base-2 was capable of getting near line rate.

What confused a lot of people was a theoretical study that showed a possible floor to the upper limit at 1/e (37%) utilization. A lot of people misread that, and assumed this was a best-case saturation point.

That myth was put to bed in 1988, with a report from DEC's Western Research Labs (DEC-WRL 88/4; Google). And that was still in the days of CSMA/CD: very few people use hubs today, so the topology has changed to almost exclusively a point-to-point model.
stephencooke 12/5/2012 | 4:08:26 AM
re: QOS Fees Could Change Everything Spelurker,

"Packet pushing boxes may use, say 10Gb/s on their I/O, but almost always have a significantly faster switching fabric. (A POS interface is serial, but while a router's internals might run at only 500MHz clock rate, they're passing data 32bits wide, giving them 16Gb/s of processing capability)"

Thanks. My question is can the routers, any of them, run at 100% utilization on all of their ports, all the time?

"In a core box, forwarding is 100% hardware, and usually deterministic, because all possible forwarding decisions have been set up based on IP destination + policy constraints."

Assuming the above, why is it not totally deterministic? Regarding your comment on all possible forwarding decisions...how can this be true when it is possible to physically move IP addresses, not just to different stationary equipment but to mobile devices (ie: an address could theoretically move from one port to another, to another, to bypassing that router all together)?

"So with today's IP boxes and LAN/WAN topologies, the effective line utilization gets MUCH higher than we were all taught was possible during the 10base-T days."

Cool. I was under the impression that the 10base-T max utilizations were on the order of 10-15%. What do you feel the current ones are?

Thanks,

Steve.
spelurker 12/5/2012 | 4:08:26 AM
re: QOS Fees Could Change Everything > can the routers, any of them, run at 100% utilization on all
> of their ports, all the time?

It's a statistical game. Rarely do traffic patterns line up so that happens. (If a highway is 4x as wide as an access road, does it have exactly 4x the traffic in each direction?) But if you throw the bandwidth at them they can handle it. Weird congestion effects do sometimes show up between ~95% and 100% utilization. Since today's links are effectively all point to point, the problems are derived from desired output bandwith temporarily bursting over 100%, and bouncing back down because of individual TCP sessions backing off.
Typical link utilization varies per location.

> Regarding your comment on all possible forwarding decisions...
> how can this be true when it is possible to physically move IP addresses,

The forwarding tables are actually pretty static.
IP routes are advertised throughout the internet, and when they get far enough from the source, they are summarized (rather than getting separate routes to Boston & New York, everyone west of Buffalo gets told to head east on I-90 for all destinations in the Northeast). Since all forwarding is done until the next hop, any one router only needs to know how to get to a router which it knows is closer to the destination. So each router normally knows all it needs to know to route a packet anywhere in the world.
However, if a nearby link fails there is a race condition between the failure event and the forwarding table update, during which time, data could be misdirected.

Mobile devices usually have all their data tunneled through a common location, limiting their exposure to routing problems. I do not know enough about the cell network itself to comment on what pitfalls lie between the access network and the metro.

spelurker 12/5/2012 | 4:08:26 AM
re: QOS Fees Could Change Everything This is a bit of a distraction from the main topic, but...

> The highest clock rate in transmission equipment and routers is the line rate of the interfaces.

This doesn't really apply. Packet pushing boxes may use, say 10Gb/s on their I/O, but almost always have a significantly faster switching fabric. (A POS interface is serial, but while a router's internals might run at only 500MHz clock rate, they're passing data 32bits wide, giving them 16Gb/s of processing capability)

> As long as the on-chip memory has been setup (either
> by recent previous packets to the same address, or via a
> connection set-up function) and the re-directing can be entirely
> done via H/W alone, the line rate can be maintained.
> Once S/W gets involved ... the actual data throughput falls below the line rate.

In practice, software does not need to get involved.
In a core box, forwarding is 100% hardware, and usually deterministic, because all possible forwarding decisions have been set up based on IP destination + policy constraints.
In an edge box, new connections generally get learned during the IP connection setup. Most protocols need to do some handshaking to start up and are moving slowly anyway. Especially TCP. By the time the IP sessions are ready to use their full bandwidth, the edge boxes are ready for them.

So with today's IP boxes and LAN/WAN topologies, the effective line utilization gets MUCH higher than we were all taught was possible during the 10base-T days.
paolo.franzoi 12/5/2012 | 4:08:27 AM
re: QOS Fees Could Change Everything
spelurker,

That method works until the moment video is deployed. At that moment, the DSLAM becomes a congested element in the worst case. Thus, all those fancy deep inspection boxes occur after the congestion.

seven
spelurker 12/5/2012 | 4:08:27 AM
re: QOS Fees Could Change Everything > Given that the carriers are selecting the home gateways that
> they deploy (I am assuming here that these gateways will
> support DSCP), this will also imply that the applications
> will also have to support DSCP to make use of this capability

Actually, that is not stricly true. Companies like Sandvine and Ellacoya provide service providers with boxes that tag packets with QoS markings based on 'deep packet inspection'. (this occurs on the network ingress, just after DSLAM/CMTS processing) No service provider in their right mind is going to trust the PC in someone's home to properly identify data streams.

> If the entire network supports [Differentiated Services]
> won't the general Internet traffic be assigned a default
> (ie: low) priority level?

Yup. That's the way the network works today. The RBOCs are (supposedly) now offering a way for 3rd parties to make use of the prioritization, which should increase the percentage of traffic which is prioritized. Therefore the BW available for everyone else should shrink proportionately. It actually shouldn't be too much of a problem, since the premium traffic will be there, whether it is prioritized or not.
stephencooke 12/5/2012 | 4:08:27 AM
re: QOS Fees Could Change Everything Spelurker,

"No service provider in their right mind is going to trust the PC in someone's home to properly identify data streams."

If the carrier has implemented an SLA with the consumer regarding the amount of traffic they are allowed at a given priority level (and presumably the price they will pay for exceeding this level), why would they care? Presumably the DSLAM, or more likely the home gateway, would be doing the policing of the SLA...?

I guess this brings up another point: ownership of the home gateway. If the consumer owns the home gateway then the only way forward is as you have described. If the carrier owns the gateway it can do the policing and prioritization prior to the traffic hitting the network. IMO it is better for the carrier to own the CPE for the above and other reasons.

Steve.
Mark Seery 12/5/2012 | 4:08:28 AM
re: QOS Fees Could Change Everything Steve,

Thanks for your response, that clears things up. I am not an expert on end station performance, but your points seem plausible. In terms of router performance, I am aware of some designs where all the traffic forwarding is done in hardware (and ternary CAMs have enough capacity for all network prefixes). I understand your point about the variability of implementation.

In terms of me and PO, all you really have to know is the the following:

There is a natural tension in many designs and architectures (technology, political, etc.) between maximum system (in this case network)evolvability (decentralizing traffic mananagement and services to the end points) and optimizing (economically, business model, etc.) a network for any given set of assumed requirements and services - this has been clearly stated even by Dr. Reed (co-author of end to end argument) who would appear to be an engineer as opposed to a sectarian. The former also requires a trust model (everyone uses TCP, new congestion management proposals for UDP, or UDP+an application level capability that is "fair" to TCP traffic) while the later often assumes that the trust model is not scaleable /trustable and therefore it is up to the operator to embed certain protections and mechanisms within the network (for example QoS) - both as a function of performance and of business model / value proposition. These are two different and respected ways of looking at communications networks, and both philosphical threads can be seen inside and outside the Internet community (see for example comments from Mr. Postel in IEN2 about how TCP should just be a host end to end congestion management, that led to the separation of TCP and IP). With respect to the latter model, it should also be noted that one end to end technology like MPLS is not required (though that may provide operational efficiencies) because per hop / per service / per flow / per aggregate behaviours can be implemented with any number of technologies. I expect that both models will exist together for a long time to come, even if in some cases one is an overlay on the other; I also believe that good architectural thinking can lead to a degree of network evolvability even in the latter model.

all the other "names" are just two people pretending they know what they are talking about and can talk in more precision than is actually likely IMO given the non uniform nature of network topologies, capacities, implementations, and measurements (for example consider the problem of distributed scheduling in an even well characterized cross bar as a starting point); my advise is to not pay too much attention - nothing wrong with being an "old optical systems guy" - there are multiple respected models of "transport" each with their own strengths and weaknesses; it should be a market place of competing ideas and value propositions; not a tyranny of thought.
stephencooke 12/5/2012 | 4:08:28 AM
re: QOS Fees Could Change Everything RJS,

Who said I wished for this...? I was just extrapolating from the announcement in the source article and some points that Seven made.

"There has to be a separation of transport and services."

Consider the guy who is the VP Sales of transport in an RBOC. His CEO is telling him that he has to start generating more revenue and his personal review is based on hitting new and improved targets. His take would likely be that transport of bits is a service.

"These things come back to haunt you and there are not easy fixes here."

I am willing to be proven wrong here. Any other prognostications? I agree that there are no easy fixes once this is started.

Steve.
paolo.franzoi 12/5/2012 | 4:08:28 AM
re: QOS Fees Could Change Everything
stephen,

The start of the QoS stuff will be video networks (aka the IPTV networks). The DSCPs will be set by the Set Top Boxes.

Eventually, there will have to be a hardware key to set the QoS at a consumer site. This is to allow these to be trusted. In the direction towards the consumer, the VPN connection itself will imply the QoS.

I would expect that these data hotels will exist in several points on the network. It may not be as much traffic as you think, especially at first. Of course, as the traffic grows the revenue is there to justify any network expansion.

And yes, you have found the flaw with QoS over the Internet. Basically, until a common scheme is adopted on every piece of equipment on the Internet it will not work.

Finally, the independent ISPs add no value to this equation. The Cable companies do not support them. The RBOCs are bigger in the broadband arena than any ISP. So, it would be simpler for a company to set up a relationship with the 4 RBOCs and the 3 largest cable companies than it would be with any other provider set.

seven
rjs 12/5/2012 | 4:08:29 AM
re: QOS Fees Could Change Everything You may actually get it!

Message 141
" ..... The days of commoditization of bit transporting are nearing an end I would say. "


Imagine a US without the commoditization of gas and oil and food. Every one of these vendors wanting to be vertically integrated to get a better value.

Very often one has to take a broader look. Mixing
State and Religion makes sense at many a level, but it is really not a good idea, no matter how many arguments we hear for it.

Common Carriage laws are the ONLY logical way and
bit transport has to be commoditized. There has to be a separation of transport and services.

These things come back to haunt you and there are not easy fixes here.

-RJS




stephencooke 12/5/2012 | 4:08:29 AM
re: QOS Fees Could Change Everything Mark,

"It was not my impression that IP/Ethernet equipment could only operate at 30% of a 10Gbps. I wonder if this is all equipment or just some? Would be interesting to hear what experience led you to this understanding and what type of equipment (if you don't want to name specific vendors I would understand)."

I can't put names to this but here is a 50kft pass at it:
- your 10/100MB/s network card in your PC has an elastic store that buffers between the main processor card and the LAN interface. The chip that interfaces to the line (lets talk CAT5 cable here) has the capability of running at 100MB/s continually. However, the software that runs that card (ie: the driver) uses the same clock. Can you see the issue? As long as the communications processing can be done in H/W alone, you can maintain the specified bit rate. Once any S/W has to do a memory access to decide anything the H/W has to insert wait states. Therefore the actual data transfer rate is much less than the maximum line rate.

- Telecom networks were all based on the same traffic type and bandwidth (64 kb/s). On call setup a message is sent the entire length of the transmission path to the destination, configuring the switch fabrics along its way for both forward and reverse paths. Obviously this is a 'connection oriented' system. Once the call was setup there was essentially a digital 'wire' between the points involved and all processing was entirely H/W-based so the line rate could be fully utilized.

- The highest clock rate in transmission equipment and routers is the line rate of the interfaces. That is divided by at least 8 (to get byte-oriented words) for processing. The chips in use today combine a bunch of functions on a single piece of silicon. What this means is that the chips that figure out where the byte boundaries are in a serial stream also do some of the routing function (eg: this packet goes to the card in slot x. The division of the overall routing function to pieces of H/W is totally vendor-dependent). As long as the on-chip memory has been setup (either by recent previous packets to the same address, or via a connection set-up function) and the re-directing can be entirely done via H/W alone, the line rate can be maintained. Once S/W gets involved (eg: a decision has to be made or a new address has to be looked up) or there is congestion on a link, the actual data throughput falls below the line rate.

- The firewall/router that you have in your home is likely a 10/100 box that has a processor that runs at <5MB/s and my be less powerful than a 486. Consider what happens when it is decrypting VPN data or doing deep packet inspection.

Bottom line is that comparing telecom bandwidths with actual IP data throughput numbers is not a fair fight. Unless vendors put beefier (ie: faster than line rate) processors on their boxes transmission efficiencies will not significantly improve from what they are now, as far as I can see.

Of course I am an old optical systems guy who knows relatively little about data routing or any of the names that you and PO talk about. I hope this helps.

Steve.
stephencooke 12/5/2012 | 4:08:30 AM
re: QOS Fees Could Change Everything Dreamer,

In the BellSouth area how much of the traffic of all ISPs, of any access flavour, touch the BS network do you think? If there is even a single network provider that implements QoS in the path to/from the consumer it will affect all traffic that touches it's network. The days of commoditization of bit transporting are nearing an end I would say.

Steve.
stephencooke 12/5/2012 | 4:08:30 AM
re: QOS Fees Could Change Everything Seven,

Thanks for the example/clarification.

"This is sharing the same physical port at the access layer as the customers Internet service, but will be differentiated by VLAN p-bits or DSCP. Thus, the networking equipment can apply distinct QoS to this traffic independent of the Internet traffic arriving at the same customer."

How many data hotels will there be do you think? Where would they be physically located? If the premium service is to be offered throughout the BS network area that means a fair amount of route similarity with the public Internet, at least within the BS area, and may use more than a little high speed backbone bandwidth.

"But nobody that I am aware of is planning a QoS service for access to the public Internet."

To differentiate the premium traffic so that QoS can be maintained throughout the traffic path will require that the entire network support DSCP...? If the entire network supports DSCP (Differentiated Services Code Point, BT announced that it is using this to offer 6 Classes of Service - CoS - throughout its 21CN build) won't the general Internet traffic be assigned a default (ie: low) priority level? My point is that, though this may not be announced as a 'service', it may actually be a necessary side effect of the premium service implementation.

Given that the carriers are selecting the home gateways that they deploy (I am assuming here that these gateways will support DSCP), this will also imply that the applications will also have to support DSCP to make use of this capability (eg: online gaming, VoIP, etc.). As the consumer may be able to generate high priority traffic it follows that Service Level Agreements (SLAs) will be the next step for carriers, based on various levels of traffic per priority level... Obviously this implies that there will be different price points for the different levels, and, wait for it.... Congress will revisit its decision not to tax all Internet traffic because of different priority levels. My guess would be that they would not tax the lowest priority level to keep with their current stance.

It must have been some bad coffee this morning...sorry.

Steve.
sgan201 12/5/2012 | 4:08:30 AM
re: QOS Fees Could Change Everything Seven,

People are allocated 20Kbps. But, it is not guaranteed. How much would it costs to make those 20Kbps GUARANTEED as opposed to only ALLOCATED?? My contention is it does not costs that much. I see that your believe it is hard to do...

You cannot sell something that is allocated as premium service but you can sell something that is guaranteed as premium service. In either case, it uses the same amount of bandwidth.

As a consumer, I find it unbelievable that an ISP will want to sell me all kind of premium services but it cannot even guarantee a minimum level bandwidth to Internet.

It is your belief that ISP and access provider have a choice and it is not worthwhile for ISP/ Access provider to be better bit providers. Perhaps your are right. But, somehow, in my gut feeling, I feel that the genie is out of the bottle. It will be commoditized and whoever is prepared to be a good commodity bit provider will win.

Dreamer
paolo.franzoi 12/5/2012 | 4:08:31 AM
re: QOS Fees Could Change Everything
Dreamer,

Here is the thing.

It is not worth it. The Content companies can work with 7 companies and bypass the ISPs. That will cover 80%+ of US broadband users. The cable companies are their own ISPs and do not allow for 3rd party ISPs.

ISPs add no value here.

If you bought a Frame Relay connection to your ISP it would do the proper job. Basically, your ISP would have to set this up and lease the circuit from the LEC.

And it costs a LOT more. People are allocated about 20kb/s in traditional Internet metros and backbones. Lots of new equipment would have to be built to get more bandwidth/user.

seven
sgan201 12/5/2012 | 4:08:31 AM
re: QOS Fees Could Change Everything Seven,

Come to think of it, Internet access over FR do not guarantee your bandwidth to the peering either as far as I know.

Dreamer
sgan201 12/5/2012 | 4:08:31 AM
re: QOS Fees Could Change Everything Seven,

1) To be precised, the service is called Internet access over FR. It is not normal FR service.

2) It does not cost the carrier a lot of money to implement this over existing cable modem/DSL/3G/WiMax/WISP. In some cases, it is almost zero. And, I am willing to pay $10 more per month for this service. As soon as one access provider find out they can be profitable and implement this with their existing infrastrucute, it will happen.

Dreamer
sgan201 12/5/2012 | 4:08:32 AM
re: QOS Fees Could Change Everything Seven,

By doing this, I had reduced the problem into the issue of having a guarantee Y Kbps per user to the peering point. I do not need and want carrier to prioritize based my application.

As long as one of the access provider (Cable versus DSL versus 3G versus WiMax versus WISP) provide a service that allow the guarantee Y Kbps, I have my problem solved. And, if one of them provide such a service, the rest of the access providers will have to match this or they will be out of business. End game.

Dreamer
paolo.franzoi 12/5/2012 | 4:08:32 AM
re: QOS Fees Could Change Everything
So, you don't want to prioritize your application. Nobody is asking you to.

And no you don't have your problem solved. There are constraints other than bandwidth, like jitter and latency. It is not just x bandwidth guaranteed, it is x bandwidth with y jitter and z latency.

Otherwise, it don't work.

By the way, the service you want is available today and is called Frame Relay. You can buy it anytime you want. Now, just because it is 10x the price of residential access should not concern you. It is exactly what you want.

seven
stephencooke 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything Seven,

Perhaps you should clarify what you mean by private IP networks. The last mile is used by both Internet access and any private IP services. There seems to be a physical overlap of the two IP-based 'networks' so how can one be 'bypassed'?

Correct me if I am wrong but I think you mean the Internet to be something that does not necessarily include the access portion of the carrier's networks...? This is a bit of a conundrum as there is no access to anything without that portion of the physical and logical 'network'. At one point you stated that the Internet has no storage of information, that it is connections 'across the Internet' to gain information. I think it was on the Google thread.

Could you elaborate? Thanks.

Steve.
paolo.franzoi 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything
Dreamer,

Except that the Local Access will not prioritize the traffic, so it won't work.

seven
paolo.franzoi 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything
Just as an aside, this form of QoS and private network sharing with an Internet connection is being done in the cable world. It's called Packet Cable and the Voice over DOCSIS works this way. If Vonage customers are sharing the same cable modem segment as the MSO's voice over cable, the MSO voice wins every time in terms of QoS.

Nobody seems to talk about that.

At an extended level, the use of QAM channels for Video restricts people's Internet Access over cable as well. If the QAM channels were used for Internet Access instead of video channels, there would be a lot more bandwidth available for the Internet.

Nobody seems bothered by this either.

seven
paolo.franzoi 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything
When I mean a private IP network, I mean that let's say that BellSouth makes a deal with Disney so that Disney sells it's content in a premium fashion to BellSouth. BellSouth and Disney set up a connection in a BellSouth Data Hotel.

When a customer has this premium Disney Service, instead of getting the Disney content over the Internet that traffic is redirected to this Data Hotel and the Disney content is retrieved over the Disney network. In the new buildouts, Verizon (FIOS already using DLink equipment), SBC (rumored to plan to use 2Wire), and BellSouth (rumored to use Westell) will be placing Home Gateway Routers at the customer site.

So, the content that lies on Disney's network (where the Disney content actually is) can now traverse to the premium customers over a completely controlled environment. From the content server to the Data Hotel by Disney and from the Data Hotel to the customer by BellSouth. This same content can be available in a non-premium fashion via the Internet.

Now, why would Disney wish to do this? Suppose they decided to make some shows/movies/whatever available in real time to BellSouth's IPTV customers. This would be a way for that content to be made available to these customers.

BellSouth is going to use a premium setup for it's own IPTV connection. This is sharing the same physical port at the access layer as the customers Internet service, but will be differentiated by VLAN p-bits or DSCP. Thus, the networking equipment can apply distinct QoS to this traffic independent of the Internet traffic arriving at the same customer. This traffic will be groomed at the aggregation layer of the network over to BellSouth's IPTV backbone. Internet traffic from the same customer will head to the Internet. BellSouth's IPTV network is a Private IP network from that perspective and not part of the Internet per se.

So, these connections are VPN style connections to the content. The content owners are paying VPN charges for these connections. These VPN fees are indrectly subsidized by the signup of consumers to access these VPNs. These VPNs share some amount of metro and access network with the consumer's connection to the public Internet. From the perspective of the VPN occupying that bandwidth, it restricts the flow of information from the public Internet. But nobody that I am aware of is planning a QoS service for access to the public Internet.

Nobody thought that BellSouth was putting its video service behind the public Internet did they?

You can replace Disney in this message with say Vonage or BellSouth's VoIP network or any other content provider that wishes to do something in a premium way.

seven
paolo.franzoi 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything
Nobody is talking about charging more for use of the Internet. They are talking about BYPASSING the Internet and charging for the use of Private IP networks to provide higher quality than the Internet can provide.

The Internet gets more oversubscribed as you head deeper core > metro > access. Giving you access BW guarantees means you will get the exact same service you have without it.

seven
sgan201 12/5/2012 | 4:08:33 AM
re: QOS Fees Could Change Everything Seven,

Let me play devil advocates. Let's say Disney has its own fiber and IP (non-Internet) network interconnecting all Disney servers. The Disney servers are placed in all the peering points. Why do Disney need to cut a deal with anyone?? Disney traffic will not transvere Internet anyhow. Google, Yahoo, Amazon, Ebay could have done the same thing.

By the way, the same approach can be done any place around the world by placing your servers in almost all the peering points in the world and you do not need to cut a deal with anyone.

Dreamer
sgan201 12/5/2012 | 4:08:34 AM
re: QOS Fees Could Change Everything Steve,

1) We had extensive experience from FR network that people want certain increment of bandwidth. For eexample, I would imagine if someone want a good quality VoIP line to home, they probably want Y to be 128Kbps. For video conferencing, Y = 512Kbps. Y is not arbitrary. It is defined by the distint application that consumer want.

2) Infrastructure costs

Actually, a lot if not most of my requirements can be implemented with pre-existing equipments. DSL modem is running IP over ATM, most DSLAMs in USA are running ATM. So, from DSL modem to the B-RAS it is ATM all the way. ATM VCC can be used to guarantee the bandwidth all the way from DSL modem to the B-RAS. From B-RAS to Internet Gate way router is just a few aggregate trunks running different over-subscription rate.

3) 80-20 rule
The majority of the congestion problem in the network is in the access link. It is not in the core or backbone. If I have two locations (A & B), both have guarantee Y Kbps to the Internet. Most of the time, I pretty much has Y kbps between those two locations most of the times. I had solved my 80% congestion problem. It may not be worthwhile for me or the service provider to solve the remaining 20% problem.

4) Oil and gas company is running a commodity business and they are highly profitable. There is nothing wrong in running a commodity business.

Dreamer
rjs 12/5/2012 | 4:08:34 AM
re: QOS Fees Could Change Everything One fine day Intel and AMD find that processing bits has become a commodity and now feel that they need a cut of every application that runs on their processors!

That is right ... eventhough you paid for your computer upfront, Intel/AMD could ask you to
pay a premium for every application that the
end user runs on their machines.

Most of us would be up in arms if such a thing were ever to happen ....
This is exactly what the carriers and RBOCs want.

Gentlemen use common sense --- just because RBOCs and the carriers feel that they need more returns that does not mean that we should rollover and give it.

LEAVE THE INTERNET ALONE ... it is a good thing even with all the shortcomings mentioned in the
discussion. Far better than having the RBOCs screw around with common carriage laws.
This will be worse than the fox guarding the chicken coop.

I am amazed that it even discussed!


-RJS
OldPOTS 12/5/2012 | 4:08:34 AM
re: QOS Fees Could Change Everything Most of the listed carriers asking for the QoS premium payments already have routers/switches at OLTs/DSLAMs that can provide QoS for each 'service'. Just Google the router/switch vendors. Most telcos also then have various methods to carry either ethernet or IP services aggregated over their own ATM/MPLS matrix networks with QoS. They also have ways to measure gross capacity utilization (erlangs anyone) per customer G«ˇserviceG«÷ or aggregated service.

The telcos donG«÷t have to control QoS on the CPE outbound port as the customer (CPE) manages this BW. It is only at the DSLAM/OLT router/switch, where the customer is aggregated with thousands of other customers sharing the same BW capacity that they need to segregate/prioritize their premium services.

------
The question is not whether they will use QoS, it is built in the networks. From their viewpoint they are segregating the existing best effort BW from those BW expansions needed for theirs and others new premium services. This bandwidth expansion may include port (e.g. ADSL to VDSL) upgrade and some BW in the existing access and core matrix network. So the question becomes how much they should expand their network to support premium services for themselves and others. But why should they expand their network if no one is willing to pay for expansions for theirs or others premium services? Or stated differently, what are the market projections for capacity needs for themselves and others? They just sent up a marketing trial balloon and also positioned themselves as offering BW capacity to others who wish to pay for it. Or else if they donG«÷t wish to pay for the expansion they can just use the congesting best effort BW that is currently available and for which they bought access to.

These costly expansions for premium services don't happen overnight for telcos, nor even for cabelcos. (why telcos are perceived as moving slowly) In many cases it is much easier/cheaper for them to add capacity initially, rather than later. QoS enables them to overbuild and then supply BW capacity QoS limited to what someone is willing to pay for.

As they started these network expansions, Google and others have forced the expansion size issue quicker than they had anticipated (and are accustom to) But will anyone else buy additional BW from them for premium services? The trial balloon went up so when others later need/demand it because they are congested from all the added traffic, they can argue they offered QoS BW and no one wanted it. They can then wait until they are ready to expand it again.

Many still donG«÷t understand that on a shared (matrix) network they bought access port BW and not end to end BW guarantees. They bought just the use of congesting best effort BW that is available as others are added or add traffic. As some have pointed out, they didnG«÷t even get a guarantee between the ratio of access port BW to shared BW in the access or core matrix network. But they can buy premium services end to end with guarantees through QoS on any expanded network capacity if needed.

BTW the telcos could also use QoS to reduce the existing BE BW capacity (as some have suggested) to shoe in their premium services. But I don't think that is their goal or mind set yet.

Also note that QoS allocates average and peak BW. That difference between the average and peak BW is statistically shared with others as BE in most COS/QoS schemes. This makes the shared network with QoS premium services less expensive than TDM networks. And it enables faster re-configuration (end to end) for premium services.

OldPOTS
Mark Seery 12/5/2012 | 4:08:34 AM
re: QOS Fees Could Change Everything stephencooke,

Thanks for your response.

All good questions and ones that would need to be asked.

I agree there are any number of network boundary issues that are still being discussed / resolved.

It was not my impression that IP/Ethernet equipment could only operate at 30% of a 10Gbps. I wonder if this is all equipment or just some? Would be interesting to hear what experience led you to this understanding and what type of equipment (if you don't want to name specific vendors I would understand).

I agree that all of this is a huge undertaking. But two thoughts. One, if a mechanism could be put in to assure the QoS of a Google service for example, could other business models with the same mechanisms be explored? Secondly, if access providers become very successful at being toll collectors, to what extent would they be incented to be in the value added services business; i.e. what would it say about the effective industry structure at that point?
BigBrother 12/5/2012 | 4:08:34 AM
re: QOS Fees Could Change Everything May be if you want to run 100 apps in you Intel or AMD processor and then Intel or AMD will tell you if you pay more to add more processors to run all the 100 apps you may want to do the upgrade. If you paid for a multi-cores system and Intel or AMD may said, I can give you a key to turn on more processors to increase the speed of your apps, you may want to do so. You may be paying for a dual core at first and you pay more to triple cores by just upgrading some keys. This may be the future anyway.
Mark Seery 12/5/2012 | 4:08:35 AM
re: QOS Fees Could Change Everything PO,

And perhaps if you reviewed the subject, you might realise that what you are discussing is probably inconsistent with the networking theory the access providers are proposing.

You are noodling down on one theory of networking and asserting that it is the only way of looking at things; and when you get an alternate view your reaction is to make attacks and then mark down the responses.
PO 12/5/2012 | 4:08:35 AM
re: QOS Fees Could Change Everything Mark, perhaps you could simply review what the subject is of the discussion. I merely pointed out that what is being suggested is inconsistent with the tools in use today.
Mark Seery 12/5/2012 | 4:08:35 AM
re: QOS Fees Could Change Everything PO,

You have asserted that every one is of one mind. That seems like a religious assertion to me. Once again your approach is to attack me, rather than make an argument.

Explain why slow-start was created?
Explain why NewReno is being proposed?
Explain whether or not there are any unresolved problems even with these approaches?
Explain why it is you believe that a protocol that throttles the sender is using the maximum available bandwidth, and to what degree of precision it knows what the maximum available bandwidth is, and to what degree of precision it can react to what you might be tempted to call the changing amount of available bandwidth?
Explain why a protocol that retransmits packets it assumes to be lost, and which the application in question might not want transmitted, is not actually wasting bandwidth, and therefore by definition not using the maximum available bandwidth?

In short stop making attacks, and start making an argument. and start by considering the fact that just because someone views the situation in a different way than you, does not mean they do not understand data communications. That is not an argument, that is a very weak and characterless substitute for an argument.
PO 12/5/2012 | 4:08:35 AM
re: QOS Fees Could Change Everything Mark, I've been involved with data networking in various capacities since the early 80's. Nobody who understands TCP, Tahoe, Reno, NewReno, etc would make the sort of assertions you made about UDP "not interfering" with throughput as TCP does, and about the design purposes of TCP.

If you know data networking, then you certainly haven't demonstrated that knowledge in your comments.
stephencooke 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything Mark,

Your point about the reasonable-ness of Dreamer's request and that some business manager may take it on is well taken. The business questions that need to be considered are of the flavour:

- How many people want Y bandwidth, how many want 2Y, 3Y, Y^2, 1/2 Y, etc.? If one flavour fits the vast majority of people then the FR or ATM-type approaches work best.
- What would be the cost to the carrier of implementing these types of services across the carrier's network and how is the interface between networks handled?
- What profit margin can be had for these types of traffic patterns? This often depends on the level of statistical multiplexing that the network is engineered for. It also has to do with the maximum throughput that the network can actually sustain. Current IP networks can only handle roughly 30% efficiency before packets start getting lost. Telco personnel find this really hard to digest (ie: you buy a 10GigE card you should be able to run constant 10GB/s through it right? That is what SONET/SDH and ATM do.). This is one of those telecom-quality EtherNet questions.
- Given that there will likely be a range of Y's that customers want, and are willing to pay for, how much bandwidth do I need at my DSLAMs to satisfy current and future needs?
- If the traffic crosses international boundaries how do I guarantee bandwidth and customer traffic priorities across that interface, irrespective of what Congress/FCC decides?
- If the traffic crosses international boundaries what do I have to charge to keep my revenue stream positive? Sounds like the various rates to various countries on your phone doesn't it? At least in the telephone world the bit rate/connection is fixed worldwide.

Will we get to this stage? Probably at some point. Will it happen soon? I would guess in the next 5-8 years. It will happen in individual carrier networks more quickly as video is offered.

What everyone has to see from this is that it is a HUGE undertaking to put these kinds of value-added services over the current network. The real question is whether the telcos can get 4-play and value-added services operational before their best paying customers jump to cable or not.

Steve.
Mark Seery 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything PO,

So looks like I stepped on your religion, and your adult response is to what....assert I don't know anything about data communications?

Hmmmm.

I did not assert that UDP is rate-adaptive, so please don't say that I did. BTW, I think you meant NewReno (perhaps if you want to have an adult discussion you can explain how NewReno has resolved all of Reno's problems once and for all and when we will all be using it). If you don't think that over 20 years of networking experience in many capacities, i have not seen more than one approach to end to end congestion/flow control that claimed to have solved this problem once and for all then you would be mistaken. there are many smart people working on this, that does not assure it is a solveable problem.
PO 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything "IMO, an equally valid whay of expressing what you said is:

-UDP does not interfere with your throughput in the same way TCP does.
-TCP was designed to protect a defenseless network against its users.
-in a network that is not dimensioned according to expected/predictable traffic demands, and does not control the entry and transit of flows (or aggregates), the potential for congestion is high.
"

Uhhh. No. Have you ever taken an introductory data networking course? If so, please ask for a refund.

UDP is not rate-adaptive. TCP is. Go buy a book on the matter and learn what that means. Go learn why Tahoe and Reno are of interest.

For one thing, it means that TCP guarantees that it will encounter congestion at at least one point along its path. Another way of saying that is that TCP will use the maximum bandwidth available to it on an end-to-end basis.
Mark Seery 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything PO,

IMO, an equally valid whay of expressing what you said is:

-UDP does not interfere with your throughput in the same way TCP does.
-TCP was designed to protect a defenseless network against its users.
-in a network that is not dimensioned according to expected/predictable traffic demands, and does not control the entry and transit of flows (or aggregates), the potential for congestion is high.
Mark Seery 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything There is nothing wrong with what Dreamer101 is asking for. Dreamer is simply saying that he wants the access provider to give him the option of having guaranteed bandwidth to his ISP, which he will pay more for, and then he is willing to trust that the rest of the network will work in a way that is roughly aligned with his needs. We can argue about whether the second part is realistic, but there is nothing illogical about the request.

The interesting questions are:

-what would a service provider actually charge for this; perhaps private line or frame relay provide some guides? (one being absolutely guaranteed and the other being roughly guaranteed as a function of engineering choices; and both being examples of the possibilities of engineering/business model choices).

-would a rational thinking access provider, provide this voluntarily?

-will the dynamics of competition (the rationale for less regulatory oversight) lead to this option emerging?

-what will be the result of legislation; especially if it is delayed till after the mid terms, and especially if the mid terms happen to change the make up of the congress (not asserting it will - have no idea).

Dreamers point of view is essential to understanding this debate. The debate started with the question of whether consumers should have quality access to services provided by the access provider (natively or via a controlled toll). Dreamer is turning this on its head and saying wait a minute, what I want is my own choice about what quality services I have access to. Now Dreamer may never get what he wants, but he is not the only consumer asking this very question - i.e. there is likely a market for this service, the question is whether access provider will offer it (an IMO, if there is a market for something, you can never completely discount the possibility of some business manager going after it).

It is an important question, and a question that regulators will need to take a position on eventually. Those that say the access providers have spoken with authority, are intellectually correct because SPs do own the assets - and individuals and companies should have control of what they own, but intellectually I own my pay check as well, however in practice, this appears not to be the case (for example the money the government takes out every week without my permission).

On this journey, it seems there are at least two distinct paths that could be taken: a) assert what the future will be (which would require a knowledge of both legislation and competitive dynamics; at least) or b) assert what is the right future as a way of either honing arguments for the upcoming fight over legislation, or as a way of revealing what upcoming arguments will be. I may be wrong, but I think I detect more of B) than A) - even granting A) is somewhat related to B).

It would certainly be interesting to hear more commentary on what people think the actual legislation will be (and when), and why they think congress will arrive at that legislation (not simply that a certain legislation is aligned with their view of what is right).
PO 12/5/2012 | 4:08:36 AM
re: QOS Fees Could Change Everything "what I want is for the ISP to guarantee that my entrance to the Internet is at least Y Kbps"

Your entrance to the Internet is the point at which your home network interfaces with your service provider's network: the inter-networking point. Which is within your DSL- or Cable-Modem.

Your service provider could probably provide some sort of assurance of some minimum bandwidth at the first edge device on their network, but what would be the point? TCP has its own congestion control and will regulate the bandwidth available to you based on all congestion points along the entire path. And UDP has no congestion control and operates on a best-effort basis only.

These are basic features of Internet Protocol packet transport. If you want different behaviour, you have to start with TCP and IP. And you'll have to rewrite BGP as well.

MPLS tries to run multiple segregated networks, but doesn't really address congestion on those networks except to try to adjust occupancy and capacity. And since it doesn't operate end-to-end, it doesn't address end-to-end congestion: it only helps you if you assume that your own service provider is in charge of the most congested segments of your path.
paolo.franzoi 12/5/2012 | 4:08:37 AM
re: QOS Fees Could Change Everything
Dreamer,

It is under your local access provider's control. Not your ISP.

seven
rjs 12/5/2012 | 4:08:37 AM
re: QOS Fees Could Change Everything Dreamer,
what you are asking for is perfectly
rational and reasonable and technically possible.

The problem is that as a CARRIER my agenda is not
to give you what you want!!
Bits transport is a commodity and fetches a minimum premium, but when you use it to control the content, that becomes a money maker.

No transport provider should be allowed to monitor
the bits. This is the only way and should be treated as the 11th Commandment. If the carriers are facing tough times because they are way too inefficient, then they need to apply for govt bail outs just like the rest of the guys like the big airlines and the detroit automakers!!

Atleast then the taxpayers will have a fair systems and let the internet grow and add to the goverment revenue base with which they can hand out pork to these carriers.


-RJS
sgan201 12/5/2012 | 4:08:37 AM
re: QOS Fees Could Change Everything It is possible because what I want is for the ISP to guarantee that my entrance to the Internet is at least Y Kbps. I am not looking for end to end guarantee.

That path is totaly under my ISP control.

Dreamer

sgan201 12/5/2012 | 4:08:38 AM
re: QOS Fees Could Change Everything Let me summarize what I want in a simple fashion. I pay for X dollar for my internet access. If the network is not congested, I should be allowed to burst and use more bandwidth. If the network is congested, I want a guarantee that I will get Y Kbps.

I do not want the SP to look into my Internet access and prioritize what application has higher priority. I can do this via my normal Linksys router at home.

I am willing to pay more for a greater Y.

Dreamer
PO 12/5/2012 | 4:08:38 AM
re: QOS Fees Could Change Everything " If the network is congested, I want a guarantee that I will get Y Kbps."

No such guarantee is possible, in general.

Nobody can predict where in the network the congestion will be encountered which dominates the Bandwidth-Delay Product for your network traffic.

What you want is to rewrite BGP (and possibly TCP), rather than add some sort of "QoS" to the network.
PO 12/5/2012 | 4:08:39 AM
re: QOS Fees Could Change Everything "See the following link for P-P information, there were some updated one from others. ..."

All I see at that link is a claim from a vendor trying to hype their equipment. What I need to see is a proper networking study, analyzing traffic at the edge, across the ISP, at the POP, and across the core.

The following link certainly isn't a proper study, but it does provide a much more informed view of traffic breakdown by categorizing traffic at the San Diego NAP (by bytes).
http://www.caida.org/dynamic/a...
http://www.caida.org/dynamic/a...
http://www.caida.org/dynamic/a...
unlimited 12/5/2012 | 4:08:39 AM
re: QOS Fees Could Change Everything Seven,

You might be right but I hope not. The stakes are high for all parties and it's difficult to be sure exactly how things will play out. Thus far I have argued for what I think would create the most interesting environment in the long run.

But what is actually happening? The RBOCs and Cable companies are heading into what promises to be a brutal battle. Each has much to gain at the expense of the other, so who will win, or will they both lose? That's going to be a major distraction for them. Meanwhile, there are power companies planning to get into the broadband business, there are municipal wireless broadband networks being built and cellular networks are improving their data services. The dialup ISPs (of which Earthlink is one) know they need to move into broadband and other Internet services to survive. To that end you see Eathlink starting to offer VoIP service, they are working with SK Telecom for wireless services and they are building the wireless networks for Anaheim and Philadelphia to name a few activities. I should mention WiMAX, WiFi and other developments in wireless technology introduce new possibilities for access competition. This makes a very complex and highly competitive environment that will be quite volatile. Given this as the backdrop, IMS, IPTV, and triple/quadruple plays look like elements to inflate another bubble.

Let's not stop there though. Recently consumer electronics has headed into digital media in a big way. Apple's success with digital music has got people wondering about digital video and the direction is going towards home media servers, HDTV, DVR's, and digital audio sound systems networked together over 802.11 wireless LANs. There's plenty there to occupy people's attention and finances. In other words how much competition is there for consumer's money? Will they be looking to spend more on TV services or will they look to cut the cost of some existing services? As they upgrade to HD digital cable do they finally drop their primary line and rely on their cell phone? I'm not suggesting I know the answer.

But I'll end with a thought about mobile data service. There's a lot of focus these days on mobile computing. Of course we have desktops but increasingly we have data to manage and store while mobile. The amount of compute power, portable storage capacity, and multimedia capabilties that can be packed into a handheld device is rivaling what desktops had 5 years ago. As we proceed down this path do we want seamless mobility between fixed and mobile data networks or do we just want the mobile data service? What if by the time all the fixed access infrastructure is upgraded, people find that their mobile data service is all they need and cancel their fixed line services just like is happening with voice today?

Everything is now moving so fast that industry players have to hit a moving target. Of course the quicker you move the easier it is to predict where the target will be.
stephencooke 12/5/2012 | 4:08:39 AM
re: QOS Fees Could Change Everything Unlimted,

Sorry if I misunderstood. However, the thing that you may be missing is that the carriers are scared stiff. They are losing a significant percentage of their high-revenue clientelle to VoIP providers (including their own internal VoIP business unit). Their bread and butter is walking.

Put yourself in their shoes for a minute... Their shareholders are looking for a rosy dividend picture (and their employees want to keep their jobs) which requires new sources of revenue as the old ones are leaving. They have customers who are willing to pay them more money for a capability that is available (with a relatively small amount of effort vs. deploying a new network) in their existing network. They have value-added services that they can offer as an additional revenue source but it has to be good or people won't switch from their existing cable providers to get it. They are already late to market with this service even though they have all been doing trials for years.

For that service offering, 'best effort' is not good enough for their customers to pay additional money for. I get the impression that people seem to think that the current traffic types will be wiped out completely by providing QoS. This is not the case. Those traffic types will just have to share a smaller portion of the total available bandwidth. Bottom line is that the current model, which you seem to be defending, cannot survive in its current form.

While you are correct that you pay for your current bandwidth, the law of supply and demand still holds. If there is someone willing to pay more, they will get first shot. The price of vanilla Internet access is incredibly competitive and very low margin. This is not the kind of business that carriers are interested in unless it can be turned into a means to a more profitable end. QoS on the existing network is the enabler of the services that are the more profitable end. We have been paying less and getting more in terms of bandwidth. Don't be surprised to see offers that allow you to pay less and actually get less.

Don't forget that this is a survival tactic for carriers, not a nice to have.

Steve.
BigBrother 12/5/2012 | 4:08:40 AM
re: QOS Fees Could Change Everything See the following link for P-P information, there were some updated one from others.

http://www.sandvine.com/news/p...
unlimited 12/5/2012 | 4:08:40 AM
re: QOS Fees Could Change Everything Steve,

No you do not have it right. You have created facts that I did not write or even intend. I do not expect a totally free network and indeed today I pay for my broadband connection. BTW it may help to add that I do not run any P2P applications and only some private VoIP. I only consume a modest amount of my broadband Internet connection's capacity. In all my posts I have never said I want more bandwidth or lower prices. I simply asked what do we get for what we pay. It is not quantified.

All the Internet content providers have paid for their Internet connections. Their agreements let them send and receive traffic as best-effort over the Internet. The traffic is exchanged by way of the backbones with Internet access networks (e.g. BellSouth) who will also pay for carriage of traffic. The point is the owners of the various networks that comprise the Internet (it is _not_ one network) are paid to carry the traffic.

I am not interested in IPTV, or PSTN quality VoIP or QoS. I understand that the Bell's hate the fact that all the value and innovation happens outside their control and will kick and scream about it. However, whatever they do it better have user value. Not asking for free, Steve, just value! The reality is the RBOCs are desperate for new revenues but would rather blockade their competition (the Internet) and raise prices than create real value which takes time and is risky.

As to the business case here is a thought to consider. The network infrastructure is capital intensive and likely to be low margin. The applications will be higher value and higher margin. So naturally companies want to do the latter. The trouble is that these businesses have different skill sets and very different financial performance and so will naturally be separated internally. Then the accountants begin to wonder why they can't redirect more resources to the services side because the ROI is much better. Before too long they are wanting to divest themselves of the infrastructure. If that happens then in the long term the very thing that RBOCs fear today will become reality. This is vertical integration versus the horizontal industry.
paolo.franzoi 12/5/2012 | 4:08:40 AM
re: QOS Fees Could Change Everything
This will crush Earthlink, unlimited. They will have no way to respond to the capabilities of the local access providers. Nobody cares about the Internet. People only care about what is on the other side of the Internet. If there is an alternative way that works better for them, then they will switch.

You should assume that the independent ISPs are corpses, they just don't know it yet.

The cable companies are ALREADY doing this with their VoIP offerings. These are prioritized over Internet traffic over the common access infrastructure.

seven
PO 12/5/2012 | 4:08:40 AM
re: QOS Fees Could Change Everything "Unfortunately, over the years, there have been many studies to show that P-P traffic takes up about 60-70% of the telco bandwidth, and the top 10-15% of the users take up the 90% of the bandwidth."

Where is your citation for this "statistic"?

Sure, we've all seen edge gear vendors try to hype such numbers in order to sell their stuff, but as a networking claim it's been largely discredited.
stephencooke 12/5/2012 | 4:08:41 AM
re: QOS Fees Could Change Everything OK, let me see if I have this right...

Unlimited wants a totally free IP network that supplies as much bandwidth as you can grab but is not willing to pay for it. Incidentally, 'unlimited' Internet access is not 'unlimited' in the amount of traffic you use, check your agreement. Taking the highway system as an analogy, Unlimited seems to want the government to add extra lanes on all highways without a corresponding increase in taxation or presumably loss of other services.

The people who own the network have said, fairly plainly, that they are going to implement the equivalent of toll lanes to the existing system because there are people willing to pay for this service.

Unless Unlimited can present a better business case that guarantees the services that people are willing to pay for, I can't see it happening (or more accurately, staying the way it is). As I have said before, the people with the authority to make the decision for the networks in question have made it and unless you can build your own network to compete by offering the service you want, it is moot. The better business case is to offer products and services to these carriers to help them implement their vision.

Steve.
unlimited 12/5/2012 | 4:08:41 AM
re: QOS Fees Could Change Everything "This is the well documented "IPTV" that is being rolled out. Since these share the same access network, there is the issue of congestion."

Seven,

You make a good argument for going with Cable Internet or an independent DSL provider. Hopefully this will help ISPs like Earthlink.
unlimited 12/5/2012 | 4:08:42 AM
re: QOS Fees Could Change Everything Spelurker,

"So maybe these 1000 users share 1Gb or 100Mb. This allows the SPs to economically build arbitrarily large networks. "

That would be sweet but to give you a more realistic figure for consumer broadband today, 1000 users will have more like 10Mb of Internet downstream bandwidth to share. I don't know but the subscriber connection assymmetry may actually give a better sharing ratio for upstream traffic if the Internet links are symmetrical.

What is the simplest solution that can help with fairness, because that seems to be the main concern? A fair sharing scheme would allow users to have more bandwidth if it is available but not starve the others when they need it. Any such mechanism should not differentiate between applications.
paolo.franzoi 12/5/2012 | 4:08:42 AM
re: QOS Fees Could Change Everything
unlimited,

You are still hung up on the Internet. BellSouth will shortly be offering video services that will share the access network with Internet services. This is the well documented "IPTV" that is being rolled out. Since these share the same access network, there is the issue of congestion.

These video customers pay more money for their access to receive this video offering. This video offering does not lie behind the Internet but in a Private IP network that is being built by BellSouth. To ensure proper operation of this video service offering, the video traffic that enters the access network from the private IP network will have priority over internet traffic. This will be true for all users, including the video customer (i.e. The video customer's internet traffic will traverse the access network at a lower priority than the video customer's video traffic).

This is going to happen. It is not a question of if it is going to happen. The amount of it that will happen and at what timescale are in question, but it will be 2006.

These elements of QoS, which allow this video traffic to work, are being offered to people who have content. If these content owners wish to bypass the Internet, then they can create services that are differentiated from those that can ONLY come from the Internet.

People can yell and scream about this, but unless there is a change in regulation this year (Congress will not act during an election year for sure.), this type of functionality will exist.

seven
spelurker 12/5/2012 | 4:08:42 AM
re: QOS Fees Could Change Everything This is not necessarily a measure of uplink BW (it depends on topology). The landmark P-P survey was measuring at the network peering point. (actually, a university/SP node) The traffic studied was Napster/Kazaa/Gnutella, not gaming. The relevant things that the public at large are not going to understand (or react well to) is service provider topology and P-P behavior.

An SP may provide you with a 2Mb service. They provide the same service to your 1000 closest neighbors. All of these links are dumped onto a common network. Does that network support 2Gb per neighborhood? No. These networks are designed around a model where most of the users are not active at the same time, especially in the 'upstream' direction. So maybe these 1000 users share 1Gb or 100Mb. This allows the SPs to economically build arbitrarily large networks.

The P-P (filesharing) traffic breaks that model, since it is always on and acts in the upstream direction. And it also uses very large amounts of bandwidth. (boo-hoo-hoo, the SPs will need to deal with that). Now most P-P traffic is not real time. Does it really matter if an offline movie download takes 1 hour or 2? Not very much. (not to mention the transfer is almost certainly illegal...) What DOES matter is if your Vonage packets get delayed by amounts which exceed a few 10's of milliseconds. (clicks, hiss, dropped connections) To make Vonage (or gaming, etc) work reliably without a QoS agreement, the SPs actually need to actively throttle large traffic flows such as P-P file sharing traffic. With a QoS agreement, the SP can identify the desired traffic's profile and mark it for preferred service. If this model works, the SP then can just leave P-P traffic alone, and let it grow to use the rest of available bandwidth, with the confidence that the network will not allow the preferred service to be impaired.
--------------------------------------------------
"Unfortunately, over the years, there have been many studies to show that P-P traffic takes up about 60-70% of the telco bandwidth, and the top 10-15% of the users take up the 90% of the bandwidth. So that is why QoS."

The conclusion here needs more support. What "telco bandwidth" is this. Is this a measure of the uplink capacity or a percentage of traffic volume? Same question with the second stat. This is a glass half full/empty issue. Is the problem that some users are taking what they were offered or users not making use of their allowance?
OldPOTS 12/5/2012 | 4:08:45 AM
re: QOS Fees Could Change Everything The following two posted questions raise the key issues;

seven- G«£Does anybody think say in 5 years we will not be discussing port rate as a metric but throughput?G«•

unlimited- G«£This is a glass half full/empty issue. Is the problem that some users are taking what they were offered or users not making use of their allowance?G«•

The answer to unlimitedG«÷s question is truly about trusting others sharing the BW of behaving in a civilized manner and not abusing others. My many years of working with packet networks, including IP, I learned that this is not a very good expectation, even within an enterprise.

The scenario usually goes;
A new packet link is installed with plenty of BW, by request of the G«ˇsqueaky wheelG«÷. The G«ˇsqueaky wheelsG«÷ gets the BW they want and the others do not initially need all the shared BW they purchased. But as time goes on those others find new applications that can take advantage of their share of BW, only to find that others are now using their share of the BW that they thought they paid for. So in the beginning the answer is that some G«£taking what they were offered and users not making use of their allowanceG«•. But later some use more than their share and the other G«£users not making use of their allowanceG«• demand that the SP/carrier provide what they bought. How can this be rectified? More free BW or QoS?

Now with convergence, my answer to 5 years from now is that we will fall back to something similar to the Frame Relay pricing model. In that model you purchased G«ˇPort SpeedG«• for high burstiness and G«ˇQoS -statistically shared BWG«÷ to secure consistent throughput for premium services. The rest is left to BE. This becomes the business/marketing answer for all those that thought more than their share was free. Unless they want to spend $$$ money to further upgrade their network.

OldPOTS

Can Bursty and Consistent users be converged?
Avoid Congestion @ a network routing point.
Queuing and Scheduling packets matters
IsnG«÷t this a Traffic Management Problem?
Just Remember G«ˇbusy hourG«÷ & G«ˇCCSsG«÷ for packets?

unlimited 12/5/2012 | 4:08:46 AM
re: QOS Fees Could Change Everything That's the right response from Google.

They already pay for connectivity. To then have to pay multiple times is rediculous. Of course under a fair scheme Google would need to make a reciprocal charge to access providers sending packets to them!
unlimited 12/5/2012 | 4:08:46 AM
re: QOS Fees Could Change Everything "Unfortunately, over the years, there have been many studies to show that P-P traffic takes up about 60-70% of the telco bandwidth, and the top 10-15% of the users take up the 90% of the bandwidth. So that is why QoS."

The conclusion here needs more support. What "telco bandwidth" is this. Is this a measure of the uplink capacity or a percentage of traffic volume? Same question with the second stat. This is a glass half full/empty issue. Is the problem that some users are taking what they were offered or users not making use of their allowance?
rjmcmahon 12/5/2012 | 4:08:47 AM
re: QOS Fees Could Change Everything Thanks for the info seven.
BigBrother 12/5/2012 | 4:08:48 AM
re: QOS Fees Could Change Everything Unfortunately, over the years, there have been many studies to show that P-P traffic takes up about 60-70% of the telco bandwidth, and the top 10-15% of the users take up the 90% of the bandwidth. So that is why QoS.
paolo.franzoi 12/5/2012 | 4:08:48 AM
re: QOS Fees Could Change Everything
BigBrother,

I have been saying in this article that QoS makes sense because of the P2P phenomenon (as an example) that you quote. If QoS is not deployed, companies like BellSouth can not offer video. Once they have the QoS capability, it makes sense to offer it to others.

But the point of this missive is to discuss your comment. You have quoted a prime example of the 80/20 rule. That a small group uses most of the bandwidth. My question goes back to what I was arguing about earlier in the thread. So what? Do they not deserve to use every bit of bandwidth all the time? I believe that what has been sold is port rate. There is a huge grey area about actual throughput coming up. Does anybody think say in 5 years we will not be discussing port rate as a metric but throughput?

I personally am not sure. If carriers (and I include MSOs here) can sell the QoS process, then that is one answer. If not, do we think that having deeper guarantees is a better idea?

seven
mr zippy 12/5/2012 | 4:08:49 AM
re: QOS Fees Could Change Everything A qualification,

While I realise that people aren't quite making this assumption, a lot of the time when this QoS issue is raised, the people for QoS seem to almost come from a position of "those other people aren't paying for what they're using, and we need to put a stop to that." If that was actually the case, then the SP will go out of business rapidly, because not only won't the SP be making a profit, they won't be even covering their costs.

with "QoS" meaning prioritising some applications over others universally, then then charging for the priviledge a la. the article.
mr zippy 12/5/2012 | 4:08:50 AM
re: QOS Fees Could Change Everything I would also want to make sure my Service Provider keeps these P2P, movie and music download types who tie up our networks off the network unless they pay for the bandwidth they use.

So you SP is giving these P2P, movie and music downloading people bandwidth for free ? Where do I sign up !

I'm guessing SPs are obviously not giving away bandwidth for free. So aren't these people quite legitmately using the bandwidth that the SP has provided them with, that these customers are paying to use ?

While I realise that people aren't quite making this assumption, a lot of the time when this QoS issue is raised, the people for QoS seem to almost come from a position of "those other people aren't paying for what they're using, and we need to put a stop to that." If that was actually the case, then the SP will go out of business rapidly, because not only won't the SP be making a profit, they won't be even covering their costs.
Mark Sebastyn 12/5/2012 | 4:08:51 AM
re: QOS Fees Could Change Everything Google just fired back at Verizon... indicating they will not pay.

http://www.networkingpipeline....
paolo.franzoi 12/5/2012 | 4:08:51 AM
re: QOS Fees Could Change Everything
rj,

The fiber comes from Corning but there are some ADC components in the network. The Video overlay equipment was originally Harmonic, now Scientific-Atlanta. D-Link does the Home Gateway Routers. Motorola the encoders and Set Top Boxes. Tellabs does most of the OLT and ONT electronics, although Motorola does some business here as well. Ciena was deployed as ATM muxes, but these are being displaced by Juniper Routers.

I have probably missed some elements.

seven
BigBrother 12/5/2012 | 4:08:52 AM
re: QOS Fees Could Change Everything There are some isp in other countries that host the xbox live in their own network. I would think people are willing to pay an x amount of $$ to see their opponents first in the online game than their oppontents see them first. Eventually they all want to have the level playing field and willing to pay a bit more for QoS.

All bets are off for QoS if the other end does not have QoS.
wap545 12/5/2012 | 4:08:52 AM
re: QOS Fees Could Change Everything I am also a Gamer and I also work for a Service Provider offering the following Broadband connections DSL, Cable modem and new FTTH/Ethernet.
Here is the quandry:
As a gamer I would gladly pay a premium to maximize Latency/Jitter free access to my Game Server (which is World of Warcraft). I would also want to make sure my Service Provider keeps these P2P, movie and music download types who tie up our networks off the network unless they pay for the bandwidth they use.
As a service provider I would love to work a deal with WOW (or Microsoft/Sony/Nintendo)in which I would deploy a very high quality connection to their central or regional server's and offer my subscribers fee based access to this premium link. I would also allow best effort links as needed at regular rates.
The ideal for all gamers is to have the Server Platform (via a Grid like network) collocated with our IP MAN Data Centers and allowing top notch links.
I am paying $14.95 to WOW now so another $4.95/mo would still keep me under $20/Mo..
I am sure WOW would love to get another $2.00/sub/mo and open up a high quality link for them.
Vonage/Google/Yahoo and Microsoft would just have t get over it and offer both a Best effort and a premium link for their Video services.
God! I hope someone would do something to get rid of the need for Real/MicrosoftMedia Video downloads which really suck.

Jacomo
Duh! 12/5/2012 | 4:08:52 AM
re: QOS Fees Could Change Everything rjcmcmahon asks:

Who is VZ buying their fiber from? Show me some growing revenues by obvious suppliers rather than more anecdotal stories and press releases claiming millions of homes passed.

and by sheer coincidence in http://www.lightreading.com/do...
R. Scott Raynovich writes on the same day:
James Flaws, Corning's vice chairman and CFO, was the lunch speaker on Tuesday... So what about optical fiber? Flaws (well, nobody's perfect) says that thanks to the FTTP buildout by Verizon, that market is growing again.

More "anecdotal" evidence: a month or so ago, Verizon had bucket trucks in the streets in the residential neighborhoods around my office. Some poles have sprouted shiny new splitter cabinets and the strands between them have shiny new splice cases. This scene has been repeated in other Massachusetts communities(that I've seen with my own eyes), as well as other states.

More "anecdotal" evidence: FiOS marketing campaign in full swing. I was at the Home Show on Sunday, and they had their live demo van in the building, and at least half a dozen marketing and tech reps on the floor and answering questions. And they were signing people up left and right. Plus billboards in and near served communities. Plus mailings and bill stuffers. Some of my friends are getting tired of hearing about it.

I don't know how much evidence you need, or, rather, how much it will take to make you stop whining. But if your interest is more than rhetorical, you could look into Verizon's most recent 10Q forms. There is a chance that you might be able to deduce something from Corning, Motorola and Tellabs's 10Q's as well, but probably not much.

This line of questioning is beginning to sound like the intelligent design zealots' "doubts" about evolution.

FiOS is for real. Get over it.
paolo.franzoi 12/5/2012 | 4:08:55 AM
re: QOS Fees Could Change Everything
madrigael,

First, BellSouth is building their video network for their own business reasons. I have no idea if it will be successful or not. But they are building it. As is SBC and Verizon. We shall see about Qwest expanding on their VDSL deployments in Phoenix and Denver.

I replied to a comment about latency as this is the issue for gamers. What happened in the 90s is very different than it is today. In MMORPGs, as an example, World of Warcraft has passed the 5 Million subscriber mark. Everquest wasn't even around till the end of the 90s and it didn't peak until 01 or 02.

However, I agree it will not be a big dollar adder. Look at it this way. Let's say that there are 100,000 WoW players in BellSouth. If all of them are willing to pay $2 more per month for their subscription for the low latency version ($1 for Blizzard, $1 for BellSouth), there is $100,000 a month for very little if any investment.

Nobody is expecting people to pay a LOT more. What we need for Broadband is services like Caller ID. Most of the profit from voice comes from these little add on services.

Since the network is getting built anyway, why not offer the capability to others?

seven
madrigael 12/5/2012 | 4:08:56 AM
re: QOS Fees Could Change Everything Brook,

First, I'm not a network guy, I'm a gaming guy, so take all this with that grain of salt. My basic idea is that RBOCs and service providers can do whatever they want, but consumers might have a surprise in store for them if all these new triple play networks are "solving" a problem that never really existed.

(Incidentally, my cell phone and satellite dish work just fine. It already gets bundled anyway. Why do I need video and voice delivered over an IP network?)

Remember Mpath and all those third party game aggregators who were going to bypass the Internet to provide better QoS for gamers in the 90s? Well, they all went out of business for a variety of reasons.

Gaming, because of the latency issues you mentioned, was at the forefront of this type of discussion long before this current round. Yet the solution you are suggesting never panned out. And that was in a pre-broadband time.

The problem was that people just got to accept slightly lower quality of service for much less money. (Because eventually these costs get passed on to the consumer...)

Gamers, the most demanding of Internet consumers, were just not willing to pay for much higher QoS. I have no doubt that a lot of what you say will transpire, I'm just not convinced that consumers won't just do an end run right through the Internet to the services they desire at a price they consider right.

I mean, does anyone honestly think that consumers are going to pay a 20% (or 5%) premium for an MP3 to download faster than cable or DSL already provide? People voluntarily set up wireless routers that substantially slow them down (hurt QoS) all the time. Performance is not always (or even usually) what mass market consumers are after.

I'm more than willing to be persuaded (would, possibly, like to be)... I obviously don't know all the ins-and-outs of the network stuff and come from an industry where content has always been considered king.
rpk3 12/5/2012 | 4:08:58 AM
re: QOS Fees Could Change Everything I think you hit it on the head here seven...

Just a few unformulated thoughts...

xDsl contracts, and likely most MSO contracts, do not offer a sustained speed to the end user, always best effort "up to"-type language.

QoS won't run through public peering

The BOC & MSO ISP products are not common carriers
at this point in time

AOL is still enormous in customer base

We are still talking as if there is a robust ISP
market (exclude dial-up and who is left?)

BOCs & MSOs will not block connections, simple
saturation of their outbound links will do that
just fine (the only customers that will notice
will be those that require streaming, most of
the popular web stuff seems to do well using caching techniques).

Again, AOL seems to continue doing well...

My speculation is that we see the BOCs and MSOs,
neither of which want to be dumb pipes, partner
with various content providers to become the new improved version of the 1995 dial-up providers.

America Online
CompuServe
Prodigy
Delphi
Genie

(we now have T, MCI and Sprint back in the big
three mass-market positions)

Which of the 5 are still around?

So the BOCs rant and rave about charging, but
they are more likely looking to partner and ultimately build alliances.

I do suspect pay game servers will pay for the latency reduction, but I do wonder if they can
take the hit of the increased bandwith pricing...
how do the Cogents of the world fare in this
environment?

Just one possible and highly viable outcome IMO.

How long until the dust settles?


rjmcmahon 12/5/2012 | 4:08:58 AM
re: QOS Fees Could Change Everything Who is VZ buying their fiber from? Show me some growing revenues by obvious suppliers rather than more anecdotal stories and press releases claiming millions of homes passed.
spelurker 12/5/2012 | 4:08:58 AM
re: QOS Fees Could Change Everything > The RBOCs access infrastructure doesn't support realtime video.
> The so-called upgrades are promises not reality.
> If their infrastructures supported video they would have gone
> head to head with cable already.

Don't lump the RBOCs all together. I live in a Verizon region -- this summer, they were laying down fiber at an amazing rate. Now they are drowning the area in aggressive "FiOS" advertising. They call my home EVERY $%#&ing NIGHT to get to me. They sent a pamphlet to my house BY FEDEX! They aren't advertising video yet, but it shouldn't be too far away -- the infrastructure is obviously in place at the edge.

> Another indicator of real actions vs. empty promises is when
> suppliers in the VoD equipment space start making money selling
> gear to telcos. From what I read, it isn't happening.
> It's all posturing and no action.

I happen to know of some deals have been inked. (I cannot name RBOC or VoD suppliers) I do not know how much money has changed hands.
rjmcmahon 12/5/2012 | 4:08:59 AM
re: QOS Fees Could Change Everything seven,

The RBOCs access infrastructure doesn't support realtime video. The so-called upgrades are promises not reality. If their infrastructures supported video they would have gone head to head with cable already. Another indicator of real actions vs. empty promises is when suppliers in the VoD equipment space start making money selling gear to telcos. From what I read, it isn't happening. It's all posturing and no action.
paolo.franzoi 12/5/2012 | 4:08:59 AM
re: QOS Fees Could Change Everything
rj,

There is no network upgrade required for BellSouth to do this. They simply offer the content owner access to the mechanisms that their own video offering will use. You have noticed that BellSouth, SBC and Verizon are all upgrading their networks anyway right?

So, for the RBOC there is 0 money to spend as they are already spending it. For the content owner and the RBOC, there is potentially more revenue. And the best part of this deal for an RBOC (and MSOs could do the same thing here) is that getting content is what locks customers in.

Now, you are right that content is an unsure business with more inherent risks. But I am glad you have determined the future of Google and other content owners. I am particularly pleased that you want to close a network up so that nobody can have access to certain facilities to make money.

seven
rjmcmahon 12/5/2012 | 4:08:59 AM
re: QOS Fees Could Change Everything Seven,

The jury is still out on Google's video download store and particularly its pay per download model. My guess is that it won't pan out and that the final model will end up being advertising based. Regardless, there's probably not enough revenue behind it to pay the RBOCs costs to upgrade the access networks such that they support realtime video. So if it works at all, it remains an internet download service where best effort is good enough for the majority of the market.
paolo.franzoi 12/5/2012 | 4:08:59 AM
re: QOS Fees Could Change Everything
rj,

That is what the search engine does. I would not expect Google to disconnect from the Internet in order to search it.

Have you noted that they (Google) have launched a video service? Do you think that if they interconnected with a local access provider and created a relationship that allowed this video traffic to be prioritized over normal web surfing that the service would perform better?
If it performed better and you could "buy" a number of videos (say this week's "24" episodes) for a small fee, that you might consider this an option. Suppose, you might want to be able to view this video in real-time commercial free? See how this might work? Perhaps Viacom might try a similar offering.

Have you noted that corporations have connections to the Internet and other IP connectivity that bypasses the Internet? Which of these services do you think is more quantifiable?

seven
rjmcmahon 12/5/2012 | 4:09:00 AM
re: QOS Fees Could Change Everything seven,

If Google bypasses the internet, where exactly will they get the content they index from? Their stated goal is to organize the world's information. Where does one find that?
DoTheMath 12/5/2012 | 4:09:01 AM
re: QOS Fees Could Change Everything Dothemath,

I've always wondered when a vertical model works vs a horizontal (or transactional one). Any opinions? Do you consider the cable companies and the cellular companies as vertical? They are the only examples from recent history for successful facilities buildouts in the US.

---
Truck-roll charges were used by some poster to justify why RBOCs have such a large employment base [and therefore have to do these kinds of revenue-grabs] and also why someone like Google could never enter this business. I disagree. Examples ranging from cars to plumbing show that there is no need for this employment base to sit in RBOCs, where it is not as efficiently utilized as it would be in small, private, local contractors. Going horizontal would liberate these employees from corporate, centralized bureaucracies, while offering customers more flexibility.

Virtual integration is easy today: just a few days ago, my car gave out on me, and I called the insurance company [which offers these services as part of the deal]. Within 30 minutes, a local tow-truck company had sent a guy to tow it. The local guy does far greater volume than what this one insurance company can justify.

Look at the SBC-Yahoo DSL deal. It is quite possible, in a few years, that the customer really cares only about the Yahoo brand. So it effectively becomes Yahoo DSL (as it already is true in Japan!) with the billing relationship with Yahoo, as part of a bundle of services.

After all, it would be perfectly easy for Yahoo or Google to deliver the traffic right at the metro/DSLAM level - in effect, they become the equivalent of the "long distance" companies [their long-haul network a mere adjunct to their extensive services operations], that interconnect with the local access networks. With cable, Wimax and 3G based access networks, it is Yahoo/Google that are in the catbird seat.

I don't see what leverage SBC or BellSouth have in that scenario. The customer is paying for DSL so they have to deliver the traffic. And the content providers politely refuse the long haul QoS value-padded services from these inefficient RBOCs and instead offer to land the traffic at the regional hubs.

That is the future that SBC/Verizon/BellSouth is really afraid of with this QoS landgrab. But I have no sympathy for them - they richly deserve their fate.
Sisyphus 12/5/2012 | 4:09:01 AM
re: QOS Fees Could Change Everything It is not an illusion - it's already a reality in architectures that are being rolled out, otherwise many services can't make it. VPNs or IPTV and several others use strict traffic isolation methods. Just because right now subscribers are used to paying a flat fee for Internet access does not mean that QoS isn't already widely implemented. Nor does it mean people would necessarily pay for different flavors of DSL access, but rather that they'd chose additional services. It doesn't make much sense to discuss QoS just as a technology anymore - it simply results from additional services being developed, or the necessity to converge existing services into the infrastructure for cost efficiency.
paolo.franzoi 12/5/2012 | 4:09:02 AM
re: QOS Fees Could Change Everything
PO,

Google is planning on offering content. That is the point. IF they want to connect to local access providers to make their content arrive at a higher quality than generic content arriving over the Internet......

seven
Krypton_Blue 12/5/2012 | 4:09:02 AM
re: QOS Fees Could Change Everything Implementing QoS for the internet is simply an illusion. Sounds good however "everyone's" traffic will simply become prioritized which will lead us to the same thing we now have ...best effort for everyone to gain access to the largest pipe available.

QoS for Carriers though is a great marketing ploy to entice users into paying higher fees for a priority they are unable to control in the overall end-to-end connection.

Carriers will market this to the masses as the propblem for their internet wait-time and that QoS is the solution for the added fee they will need to pay to be first in line. Problem is, as many have stated, everyone soon becomes first in line as your packets travel through the vast autonomous systems controlled by a myriad a different companies.

So in the end, forget about such things as spam chocking 50% of the available bandwidth today, the Carrier simply needs to create the perception of the problem and that QoS along with its fees will solve that nasty wait time for your files to download.

Marketing Genius!!!
Sisyphus 12/5/2012 | 4:09:03 AM
re: QOS Fees Could Change Everything
The fact that the "valuable content" is not part of the actual infrastructure is something that's discussed as being new, but it has always been the case in telecommunications. Every time somone used the phone network, it wasn't to talk to a MaBell operator, but rather to a person that dynamically provided "content" one wanted access to on the remote phone set.

Telecommunication operators have always made their revenue by selling "guaranteed connectivity", and the higher guarantees and connectivity, the higher the charges. That's why QoS is natural, ultimately it introduces more choices for users. Whether or not it will be implemented fairly we can just speculate about right now, but the fact it is a useful concept is pretty evident.

A lot of the debate is driven by the natural clash of interests: of course today's content providers can claim they own the value, but again, that's nothing new when it comes to communications. So what's ultimately going on is the usual posturing to protect one's turf and threaten to invade the other's: traditional service providers wanting to get a slice of the content business, and content providers claiming they can build their own infrastructure. But only cable operators (who already own an infrastructure but whose owning of content is equally questionable) and very few priviledged companies (Google and... who else?) can credibly talk about multi-$B infrastructure investments. All of them will face major challenges executing those plans. Sure offering content is new to traditional wireline operators. But oddly enough now one's talked much about how it'll affect Google that their financial fundamentals all of a sudden would look more and more like, say, SBC's if they truly roll out a major telco infrastructure.

It'll interesting how it'll pan out, but it has all preciously little to do with QoS, but rather with everybody involved trying to milk the usual % of extra revenue and margin.
PO 12/5/2012 | 4:09:03 AM
re: QOS Fees Could Change Everything "1 - Perhaps you have not been reading about Google's plans. The search engine itself is not content."

You're right. The search engine itself is not content. (But the search engine may be subject to patent law.)

But then again, I've never seen Google's search engine. I have, however, seen the results of their search engine thousands of times.

And the results of their search engine is most emphatically content. (And the results of their search engine may be subject to copyright law.)

Now if you have a point to make, I'd be interested in reading it.
paolo.franzoi 12/5/2012 | 4:09:03 AM
re: QOS Fees Could Change Everything
No rj, the content is at computers that today are beyond the Internet. The Internet is the connectivity to those computers. If people like Google, Time Warner, Disney, etc. build networks to bypass the Internet for premium applications then the Internet is not required to connect to this content.

The Internet contains 0.00 bits of content.

The idea is to offer capabilities to people to allow them to make services. If nobody buys the capabilities, then this will go away. BellSouth will be building the capabilities for its own video offering (which requires QoS). It is offering them to others. Assuming that pricing is fair and access is reasonable, I am unclear why this is a bad idea.

seven
sgan201 12/5/2012 | 4:09:04 AM
re: QOS Fees Could Change Everything
brookseven,

Because of how the network is currently designed, when the network is congested, individual users are not guarantee to get a slice of the bandwidth. But, if the network is designed in such a way that even during network congestion, each user will get X Kbps. This design plus priiritization in the customer's CPE would have make the VoIP works well. VoIP is a low bandwidth application.

Dreamer
paolo.franzoi 12/5/2012 | 4:09:04 AM
re: QOS Fees Could Change Everything
That is QoS dreamer. Get it? Guaranteed minimum bandwidth would be a QoS service. Today, it is all best effort...you get what you get and if ain't enough...too bad.

seven

rjmcmahon 12/5/2012 | 4:09:04 AM
re: QOS Fees Could Change Everything 1 - Perhaps you have not been reading about Google's plans. The search engine itself is not content.

Right. The content is found on the internet. Anybody building a network to bypass the internet kinda misses where the content resides and misses the value proposition.

2 - Who knows? Neither do they. The idea is to allow premium services to exist so that others can make them.

If the application has to subsidize the buildout, then the first order of businesses is defining the application. Hoping for magical QoS enabled service just obfuscates the fact that there is no new revenue to be had. No killer application then no new facilities with this model. And the US falls further behind the rest of the world with respect to communications infrastructure.
paolo.franzoi 12/5/2012 | 4:09:04 AM
re: QOS Fees Could Change Everything
rj,

1 - Perhaps you have not been reading about Google's plans. The search engine itself is not content.

2 - Who knows? Neither do they. The idea is to allow premium services to exist so that others can make them.

seven
rjmcmahon 12/5/2012 | 4:09:05 AM
re: QOS Fees Could Change Everything Their {the RBOCs] vertically integrated business model is obsolete.

Dothemath,

I've always wondered when a vertical model works vs a horizontal (or transactional one). Any opinions? Do you consider the cable companies and the cellular companies as vertical? They are the only examples from recent history for successful facilities buildouts in the US.
rjmcmahon 12/5/2012 | 4:09:05 AM
re: QOS Fees Could Change Everything Seven,

Google doesn't own content. They index content found on the internet.

Also, what revenue generating services is Bell South going to enable with QoS?
paolo.franzoi 12/5/2012 | 4:09:05 AM
re: QOS Fees Could Change Everything
Again, I think people have this all wrong. Networks are constrained environments. No matter what you do, there will be periods of peak utilization that congest networks. It happens every day. If you are going to make applications that depend on having bandwidth available, then you need QoS. Otherwise you will end up with people's stories (see earlier in the thread) of VoIP calls working sometimes and not working sometimes. You can have a local port rate of 1 Gb/s and still face network congestion.

Second thing that people are missing is that mixing QoS and the Internet is not going to happen. That will remain a best effort environment. These "QoS-enabled" connections will be between content owners (example: Google) and Access Providers (example: BellSouth). There will be control across the access provider's network, because the Home Gateway Routers are going to be carrier provided (at least to start with). This environment can be completely controlled and thus a quality service provided.

Yes, it is not the Internet. Yes, small companies offering innovative services will have trouble getting started. But that is a maturing business and the way it works.

And by the way, if you talk to a gamer...Latency is his/her only issue. They have plenty of bandwidth today. But to reduce the ping would be a great deal, otherwise where is the interactivity.

seven

mr zippy 12/5/2012 | 4:09:06 AM
re: QOS Fees Could Change Everything QoS makes sense - for the operator, they get to differentiate their service thereby making incremental revenues, the end user gets a faster service, the service provider (being iTunes etc) sacrifices some margin but gets better access to their customers, and can also differentiate their revenues to their customers (click here, and for another 20 cents, we guarantee faster downloads!!).

I don't necessarily think end users buy their bandwidth that way. Bandwidth is cheap enough, much like desktop performance, that they buy based on what their peak performance expectations are, rather than their average.

From part of earlier post I wrote in another thread :

http://www.lightreading.com/bo...

Firstly, I think people buy their bandwidth not based on their average utilisation, rather, what sort of performance they want for instant or peak utlisation. This is similar to how they buy their PCs, as discussed by Clay Shirky in http://shirky.com/writings/gri.... Further evidence of this is how people seem to disproportionately care so much about how quickly an application, such as MS Word, starts and is available after they click the icon, rather than caring as much about how it performs once it has started, and their using it. Humans are after "instant" statisfaction, or at least statisfaction within no more than 3 or so seconds.

When it comes to the Internet, I expect a web page I access to be displayed as fast as my broadband link operates. Accessing web pages is certainly bursty in nature, however I think it causes people to expect that sort of performance from all applications they use the Internet for, such as file transfers and P2P i.e. constant load applications. I commonly see people trying to buy the largest bandwidth broadband links to "speed up their downloads", so that a 600MB ISO file only takes 5 minutes to download.


Thinking about this a bit more, and summarising, it seems that people are willing to spend the most on the most obvious thing that performs the least.

Startup times of applications is the thing that performs the least on a desktop PC for most people today, not the operation of tha application itself, so they now judge the performance of their PC based on application startup time, not the application's operational speed.

For broadband, download speed e.g. the time it takes to download a 3 or 4 MB MP3, or for "downloaders", the time it takes to download a ISO, a TV show (legallitiess ignored), are the things that have the least performance, and so people are willing spend on them. Latency is of course important, however I think it has become less so as the bandwidth of most broadband services today can easy meet the latecy requirements of most applications where it matters.

I'm not really saying that QoS isn't something that needs to be worried about anymore, rather, that I think that it is not the primary thing people are willing to spend a significant proportion of their "broadband budget" on achieving. It seems that a lot of QoS problems I hear of people having with existing broadband services are resolved by prioritisation on the CPE, hence you tend to see "VoIP optimised" CPE these days. To get acceptable QoS for these applications, nothing is done on a per customer basis by the upstream ISP.

In that context, where throughput seems to be becoming the dominant thing people care about, could the RBOCS be sued for artificially degrading a service so they can squeeze more revenue from there customers ? (Bearing in mind I'm in Australia and don't have any idea how trade practices laws work in the US)
PO 12/5/2012 | 4:09:06 AM
re: QOS Fees Could Change Everything @Stephencooke, "If you feel that these companies should spend billions to upgrade their networks so that they can get zero return on investment...stay out of management. The carriers have plainly stated that they are not in the business of being 'dumb bit transporters'. You can argue this until you can't type anymore but those with the jurisdiction to make the decision have done so and the rest of us have to deal with it as fact."

Then we need to get these carriers out of the business of mixing their phone and internet operations. This is already supposed to be required by federal law, but apparently it isn't working.
PO 12/5/2012 | 4:09:06 AM
re: QOS Fees Could Change Everything What does QOS mean to you? Does it mean low delay? Well, I don't see anyone offering to raise the speed limits for light in a fiber optic cable, as they might on an Expressway.

Does it mean low packet loss? But packet loss is a feature, not a bug, in TCP/IP networks, serving to identify the upper limits of the available end-to-end bandwidth.

Perhaps people are talking about "overbuilding" networks with optimized connectivity. But unless they're going to prevent other traffic from taking advantage of those networks, what good does that do? And then all they're really talking about is working to give "normal" traffic worse service.

Often, when people talk about QOS, they're talking about contention resolution algorithms: how does a forwarding point decide which of two or more packets contending for the same output port should be forwarded first, and which should be "delayed" or dropped. And how much delay might a packet incur, if it is delayed?

But where does this contention occur? It can occur on your local network (e.g. your in-home LAN). It's likely to occur at your access link (e.g. DSL modem). It can occur crossing your ISP's network (although for reasons we'll only touch on below, that's unlikely). It's likely to occur when your ISP hands traffic off to the next Autonomous System in the forwarding path, and again at other network hand-off points. And it's likely to occur at the destination "access" link.

At a high-speed interconnection point, even if a packet is delayed, it's unlikely to be delayed for very long. Packet warehouses are expensive and provide no value for anybody (despite the plan of at least one equipment vendor to build such atrocities in the past).

How many of those congestion points does a service provider have control over? You only control "all" of them if you've set up a virtual circuit between the two endpoints, which requires expensive gear all along the way. That may be fine for some specialized applications, but not for the masses.

For most residential users accessing popular web pages, the Bandwidth-Delay Product is going to be dominated by either the link speed at one end of the connection (perhaps divided by the popularity of either end of the connection, if there are multiple simultaneous connections).

So what is QOS going to get you? The opportunity to have the same decision made with everyone else's "high-quality" traffic on that same congested access link. That, and a higher bill.

Why is a packet unlikely to incur congestion while crossing a network, but instead at a network boundary point? Because this is simple economics: the edges of the network represent high expenses (NAP charges, access links, etc). If congestion is occuring in the core of your own network, those edge expenses are being wasted. We could go on about this, but we'll leave it at that for today.
DoTheMath 12/5/2012 | 4:09:07 AM
re: QOS Fees Could Change Everything ... and I do belive DSL can be (if not it already is) very profitable at $20-25 for 3 Mbps, and in for 50 Mbps VDSL, something already common in Japan.

The equipment is already dirt cheap, and getting cheaper all the time. The chinese vendors are willing to deal.

What we need in the telecom business is a Dell, who accepts that kind of economics, and makes the business model work within those parameters. What we are getting instead is old IBM thinking.

Yes, jobs will be destroyed on the network side, but it will lead to an explosion of job creation on the services/applications side.
DoTheMath 12/5/2012 | 4:09:07 AM
re: QOS Fees Could Change Everything RBOCs are inefficient, even accounting for the truck-roll etc. Cars require far more customer touch points than networks ever will, yet auto companies don't employ the salesmen and auto-mechanics on their rolls (think union scale wages ...). RBOCs could easily franchise out the customer-touch-point part of the equation, and let the competitive market provide that.

So don't assume that a future RBOC replacement/alternative will operate the way current RBOCs operate. Their vertically integrated business model is obsolete. They are foisting that high cost business model on the rest of us, refusing to reform and change.

At one level, they are not really scared of the internet and VOIP as technologies; they are scared of the business model challenges, as they should be. Liberation will come only when the tired old Bellheads running these companies give way to a fresh-thinking younger generation.
rjs 12/5/2012 | 4:09:08 AM
re: QOS Fees Could Change Everything Stephen:

I do not believe that the RBOCs are entitled to any revenues. This attitude of entitlement stemming from the monopoly days is the exact thing that is the root of all the problems the US telecomm is currently facing.

-RJS

"Actually the point that I am making is that the fees that the RBOCs charge for Internet Access is not sufficient to make up for lost voice revenues via VOIP (from any carrier including themselves)."
rjs 12/5/2012 | 4:09:08 AM
re: QOS Fees Could Change Everything Most network access in India is based on
a certain amount of GigaBytes a month. Note that
in that regard the internet is really not free.
The person who is downloading P2P porno will
sooner or later have to pay for the bandwidth.

This is a nice way of keeping the telcos out of services. QoS fees is just a red herring and a veiled attempt by the RBOCs to monitor traffic and give preference to their services.

-RJS

paolo.franzoi 12/5/2012 | 4:09:09 AM
re: QOS Fees Could Change Everything
Wireless will never have the fastest connection rates. At the end of the day, Shannon and Nyquist will prevent that. Fiber has the lowest noise and Copper has significantly lower noise. Physics is physics.

I think people misunderstand the whole QoS thing. Unless there is a way to control QoS and enable QoS, then premium services will ONLY be available to the local network provider. Unless there is control ALL packets will be marked as expedited. At that point, all QoS disolves. So, BellSouth is trying to promote a method that allows for new services. Nobody is going to crank down the Internet per se. But if you are paying good money for an application, you would like it to work. You would like it to work no matter how much porn your neighbor is downloading.

seven
stephencooke 12/5/2012 | 4:09:09 AM
re: QOS Fees Could Change Everything Dreamer,

In existing areas that currently have no/little phone service, wireless is definitely the way to go. However, for anywhere with existing infrastructure (ie: cabling) the people who installed that infrastructure justified the investment based on several decades of payback. They aren't going to be too happy to just let their investment go away by allowing cheaper players to come in without paying their dues.

The business case in telecom today is finding innovative ways to help carriers get more return on their infrastructure investment (ie: twisted pair). New flavours of DSL are part of the solution but carrier business models have to change as well as the technology. The Home Gateway Initiative (HGI) is the type of platform that will need to be built with multiple access technology options (fibre, wireless, twisted pair, others?). Value-added services are definitely the way carriers have to go and QOS is just a part of the overall solution.

Please note: "solutions" that do not raise the bar, compared to competitive offerings, are a waste of development time and money. The capabilities that carriers will need to invest in are those who leapfrog competitive offerings and have a lifetime of at least 10 years.

Steve.
sgan201 12/5/2012 | 4:09:10 AM
re: QOS Fees Could Change Everything 1) QOS on WiFi 802.11/b/g was a big deal until peopel start seeing that 802.11n is going to deliver 100Mbps and more.

2) Access is all about cost. And, to have low cost, you need to have the volume. The largest access networks are being build now in India/China and other emerging countries. It is not in USA. So, if you to know and see what kind of next gen access network will be built, it will not be in USA. USA will be the follower.

3) So far, in emerging countries, for access, wireless is the only cost effective option.

Dreamer
TheMuffinMan 12/5/2012 | 4:09:10 AM
re: QOS Fees Could Change Everything QoS makes sense - for the operator, they get to differentiate their service thereby making incremental revenues, the end user gets a faster service, the service provider (being iTunes etc) sacrifices some margin but gets better access to their customers, and can also differentiate their revenues to their customers (click here, and for another 20 cents, we guarantee faster downloads!!).

This QoS fee is a red herring since nobody is going to want some competitor access to their customers on an even QoS playing field.

It won't be their compeditors, it will be their corporate customers, ie, eBay's or Google's, and interconnect agreements will take care of the inter-carrier $ just like they do today. Getting the technology to do it will take a bit more time, but it'll get there.

The bidding war for new RF spectrum is going to be huge. The RBOCs and MSOs are going to be bidding against the new generation disruptive technology companies. The name of the game is to sell services and you have to own your own access network to succeed.

I don't agree here, I doubt we'll see a re-run of the bubble period 3G spectrum prices, and as someone has already pointed out running your own access networks is a difficult thing that internet companies have no experience doing. The real name of the game is to sell compelling services and get someone to pay for the bit transport, which Apple etc have done very well so far. Its going to be a lot cheaper to sacrifice some margin to pay access providers than build your own access network.
stephencooke 12/5/2012 | 4:09:10 AM
re: QOS Fees Could Change Everything DoTheMath,

Actually the point that I am making is that the fees that the RBOCs charge for Internet Access is not sufficient to make up for lost voice revenues via VOIP (from any carrier including themselves).

I don't think anyone would argue that the RBOCs are inefficient. However, if you feel that Google has a clue when it comes to being a regulated carrier, I strongly disagree. The long haul business is the easiest telecom business to get into. Local is the hardest. MCI did the long distance play and became one of the largest carriers incredibly quickly. They were never really able to penetrate the local in any significant way though they had a substantial revenue stream from the LD side.

Google is a big, relatively efficient machine, no argument, but their business does not charge consumers for essential services as regulated carriers do. That is a whole ballgame in itself that would require probably 10X the current staffing levels for customer service alone, not to mention E911, installation technicians, etc.

Could Google do it better? I would certainly hope so. I'm even sure that the RBOCs, if they were allowed to start from scratch, would be more efficient than they are now. Like I have said before, the RBOCs will fight the hard fight to survive and their resources go far beyond technical. These boards have been forecasting their demise for many years and they are still enormously powerful.

Steve.
ip_power 12/5/2012 | 4:09:11 AM
re: QOS Fees Could Change Everything One of the arguements the RBOC make is that these other services will affect thier Video Deployment - This is simply because they will not follow that with has already been layed/proved out by many Rural ILEC - RBOC have to do it thier way.

The other arguement is these services are also offered by them and effects the RBOCs on ability to deliver a good quality service - again I reference the above.

Finally, I aggree - I have seen it first hand and while working for one, the Revenue is and has been handed to them for to long with no real competition.

This is the late 90's and early 2000 CLEC game that the RBOCs are trying to do once more. Make it hard - if not impossible for the compitition to compete. The US need so wake up and look at the rest of the world. Becuase we the users have and we are tired of it.

They need to Focus on the market that is really hurting them CELL and MSO. Most have the investment in the Cell side so that is hidden revenue that they also get interconnect T charges for (subsidized). So that leaves the MSO, this is like the play of Cisco and MS with Business VOIP, niether really wants to fight each other so go after the others.

This is a bunch of bull and I hope the FCC steps in and stays the course of IP services is nonregulated and can not in any way be filtered or limited.



alchemy 12/5/2012 | 4:09:11 AM
re: QOS Fees Could Change Everything DoTheMath writes:
I disagree. The RBOCs are fundamentally inefficient. Even today, after all their slimming down, what they do [network management and operations] can be done with one fifth or one tenth the number of people on their rolls.

The RBOCs and MSOs have that huge headcount because they have to make residential service calls and maintain their wireline access networks. Sure, you can streamline it some but a truck roll is going to cost the company $100. If you believe that a Google can sell stock to buy licensed spectrum, build a wireless access network, and offer residential triple play product that my techno-phobe mother can self-provision, I guess you should load up on Google stock since they will rule the world.

The bidding war for new RF spectrum is going to be huge. The RBOCs and MSOs are going to be bidding against the new generation disruptive technology companies. The name of the game is to sell services and you have to own your own access network to succeed. This QoS fee is a red herring since nobody is going to want some competitor access to their customers on an even QoS playing field.
DoTheMath 12/5/2012 | 4:09:11 AM
re: QOS Fees Could Change Everything Steve:
Aren't you making an assumption that the current level of fees the service providers are charging for internet access [both at the consumer level as well as at the level of web sites like Google] is insufficient to fund their investments.

I disagree. The RBOCs are fundamentally inefficient. Even today, after all their slimming down, what they do [network management and operations] can be done with one fifth or one tenth the number of people on their rolls. It can be seen in how Google/Yahoo kind of players conduct their services operations - and Google is even in the long-haul business now. So what the RBOCs are asking for is another way to charge for their inefficiency.

DoTheMath





stephencooke 12/5/2012 | 4:09:11 AM
re: QOS Fees Could Change Everything Dreamer,

Good point. What a lot of people tend to forget is that big companies, when threatened, bring out a vast arsenal of weapons that have nothing to do with technology (eg: lobbyists, slow implementation of access agreements, frivilous lawsuits, etc.). Unless these new access providers have significant non-communications related revenue sources and an existing (or incredibly cheap lease) core network (eg: cable providers) that they can deploy in the fight it won't last long.

At the end of the day it may happen that some of the 'legacy' carriers die but I have serious doubts. Don't forget that people such as those on these boards are in the business of keeping them in business as there are big numbers involved. Unless they suddenly can't provide the cash to pay for technology that many of us would be happy to supply then I see changes but not binary levels (ie: they live or they die).

Steve.
sgan201 12/5/2012 | 4:09:12 AM
re: QOS Fees Could Change Everything Steve,

Your assumption here is we will always build access network the same way and only RBOC and PTT can build the access network. If there is a paradigm change and your assumption no longer hold true, all bets are off.

Dreamer
stephencooke 12/5/2012 | 4:09:12 AM
re: QOS Fees Could Change Everything Though I applaud those who maintain the ideology of a 'free' Internet I feel that reality has to come home at some point. While the Internet only carried file transfers, email, and websites the model at the time was sufficient and worked fine. With the advent of VOIP you are now stepping into the domain of, and taking significant revenue away from, some of the largest companies in the world.

If you feel that these companies should spend billions to upgrade their networks so that they can get zero return on investment...stay out of management. The carriers have plainly stated that they are not in the business of being 'dumb bit transporters'. You can argue this until you can't type anymore but those with the jurisdiction to make the decision have done so and the rest of us have to deal with it as fact.

So, given that RBOCS, PTT's, etc. have refused to lie down and give everyone free reign on their networks we see that they are going down the route of charging for performance. Is this a surprise? It shouldn't be. Instead of denying that this should happen the prudent way forward is to figure out how to help them make it happen in the best way possible so that everyone gets something from the pain that will occur in implementation.

Why is this happening? For those who have not had access to the telecom world for the last decade or so, carriers have been looking for the silver bullet technology to combine their multiple, disparate, non-mutually compatible networks onto. Recently they seem to have been forced to make that decision in favour of IP. They could have chosen IP long ago but they felt that it was never a telecom-quality technology. It still isn't. However, customers are voting with their feet (ie: are walking away from carriers) for it so their hands are being collectively forced.

The network on which the Internet was 'successful' is no longer the current network. As more services are kludged so they can be made to work over IP, the more the demands on the processing power of that network and access to that network are changing. Once the decision was made to put voice traffic, the life-blood of telcos, over the net, the end of 'free' Internet was in sight. If you take any set of huge multi-national, multi-billion dollar companies and tell them that their main revenue product is now to be given away for free, prepare for serious changes to the status quo.

There are two ways to look at this:

1) Aw CRAP, someone's traffic is going to get priority over my P2P video download...CRAP!!! or
2) This is a major opportunity to provide equipment and services to those huge companies.

Take your pick. Its an attitude thing.

Steve.
sgan201 12/5/2012 | 4:09:13 AM
re: QOS Fees Could Change Everything Let's just break the network into 3 pieces:

Customer to the CO -> Access

CO to the hub/Peering point -> Edge

Between Peering points -> Core

Yes, you can run a separate CORE network from Internet. But, majority of the QOS problem is at the Access and Edge. It is very expensive to have separate access and edge network to guarantee QOS. How can this be done?? Is the carrier going to control the CPE so that traffic can be prioritized?? There are a lot of $$$ involve here with little guarantee return??

I cannot see how the QOS can be deliverred?? What am I missing here??

Dreamer
robertmb 12/5/2012 | 4:09:13 AM
re: QOS Fees Could Change Everything The very reason for the growth and success of the Internet is that it does NOT distinguish between classes of service. Everyone's traffic is equally important. While there are important technical and performance reasons to avoid latency, the idea of adding more fees for unimpeded passage is equivalent to holding hostages for ransom.

Another example to illustrate the point would be similar to charging extra money to sit in the first class section of an airliner with the promise of getting to the destination before all the other passengers. We all realize this to be a bogus premise, unless the non-premium passengers are left behind because of overbookings. Kind of makes one wonder why priorities are focused on "premium charges", instead of addressing the issue of inadequate facilities to handle the traffic. Do the RBOCs have their own version of "Moore's Law", or are they still clinging to old business models?
paolo.franzoi 12/5/2012 | 4:09:14 AM
re: QOS Fees Could Change Everything
Dump the Internet. It will never have QoS. Think beyond the Internet. The future is there.

seven
unlimited 12/5/2012 | 4:09:14 AM
re: QOS Fees Could Change Everything Seven,

Thanks that's what I wanted to know.

As I suspected to you every thing is a telephone network.

Turn around the future is the other way!
paolo.franzoi 12/5/2012 | 4:09:15 AM
re: QOS Fees Could Change Everything
Wrong. You keep thinking Internet.

Take the content providers - bypass the Internet and connect to the top 50 access providers. There are these places called "data hotels" where such meeting points are made available. Perhaps you should look into them.

Get it now? Leave the ISPs et al for best effort traffic. All this QoS traffic is going around the Internet. That is why companies like Google are building their own network. They would like one that works all the time.

seven


Service providers have never proven that customers have gotten what they pay for. Remember my note about conditioned lines. It was up to the customer to prove that they were NOT getting the service they paid for.

You just have to live in your little box. I guess anything else is too hard.

seven
unlimited 12/5/2012 | 4:09:16 AM
re: QOS Fees Could Change Everything Seven,

In the Internet world there are bi-lateral peering agreements. But so what? How many service providers times how many access providers? How much complexity is that exactly?

Sorry to the 20% of you that cannot get QoS but have to subsidize service for the 80%.

Now can you explain how service providers can prove that their customers are getting what they have been charged for? Any solution to that problem will add yet more complexity. Now I understand that phone people like as much complexity as possible and as users like to pay extra for the priviledge. But hopefully there are more rational people out there.
paolo.franzoi 12/5/2012 | 4:09:18 AM
re: QOS Fees Could Change Everything
Yes, I would expect bi-lateral arrangements with say the 50 largest carriers. Gee....I wonder what happens in the phone world. Why guess what there are bi-lateral agreements with the large carriers.

By the way, those 50 carriers represent 80% of the opportunity.

Try to grasp the point. You are sold lots of things that you don't know what you are getting. Please describe what "organic" means for foods. Does it mean just that no pesticides were used or does it mean that no human made means were used at all? Does no human made means include tools? You are really being quite facetious here.

I suggest if you think that all the marketing buzz and the metrics that things are sold under are measureable in all cases then you are insane.

seven
paolo.franzoi 12/5/2012 | 4:09:20 AM
re: QOS Fees Could Change Everything
You clearly are making assumptions about what is going to occur that is not evident.

Let's use Google and BellSouth. Google puts one of their PoPs and co-locates it with BellSouth for some premium service. Google pays for this connection and both BellSouth and Google work to sign people up.

Do you think Google will not get an SLA? Do you believe that customer complaints will not be monitored/examined? If you were Google and all your customers were up in arms, would you continue with the connection?

You need to get real. If you want to know what your bandwidth is, lease a DS3 to the Internet. Are you willing to pay for that or are you willing to pay $25 a month for Internet service and expect the same level of service as Enterprises paying $1000s per month. I am suggesting that you want a lot more than you are willing to pay for.

seven
unlimited 12/5/2012 | 4:09:20 AM
re: QOS Fees Could Change Everything Seven

"Let's use Google and BellSouth. Google puts one of their PoPs and co-locates it with BellSouth for some premium service. Google pays for this connection and both BellSouth and Google work to sign people up."

I need to get real, whereas it's totally rational for Google to make special arrangements with BellSouth for QoS? What about Earthlink, Verizon, BT, Comcast, Wanadoo, ... Let's see how many broadband Internet access providers are there?

On the almost interesting topic of access bandwidth try to grasp the point. Access providers could sell broadband as "Quite fast", "Excitingly fast", "So fast you won't have time to blink". But they don't. They dimension it and nobody knows how, least of all the consumer. Support services were not mentioned. It seems like fairly standard practice in the world of retailing to explain to users what they are actually getting. I suggest if you find this point challenging you forget about it and worry about some more interesting aspects of the topic.
unlimited 12/5/2012 | 4:09:20 AM
re: QOS Fees Could Change Everything Seven

>>You can only buy a different bit rate at the access port. I could give you a GigE interface and put a router with a T-1 connection to the Internet and you would pay for the GigE. You do not pay for a bit rate connection to the Internet (in general and in residential cases). For Enterprises, there are all types of equipment to monitor SLAs.

Yes. So you are happy with this? When an enterprise buys service, they know what they are getting. Do consumers? Of course not. Just because it is done this way does not make it desirable. I don't think you are suggesting that consumers put in SLA monitoring equipment so I'm not sure what your point is.


>>The reason you are not getting bandwidth measurements is you are paying for the clock rate that the line is wiggling. That's it.

Yes sort of like buying a 100A circuit with a 1A fuse feeding it from the power company. Again for consumers this would seem to be false advertising. Get your 3Mbps Internet connection that by the way may only gives you 10Kbps on occasion.


>>You would only pay for the QoS indirectly. The application provider would pay for the QoS service and you would pay the application provider. Thus, it would be in the best interest of the network provider to provide a quality service that people wanted to subscribe to. This opens up a revenue stream. Without this, there are all types of applications that will never exist.

No one has yet explained how to get around all the pitfalls for this method. I have been trying to point out, in fact, that what you pay for becomes even more unmeasurable under this scheme. Actually it is quite brilliant. Charge providers for priority delivery of their traffic to customers. That ensures that customers don't know they should be getting better service and providers can't check that they do! You realize that in most cases the providers will have no physical connectivity with the access providers delivering their traffic. Packets will have to be matched at the borders and marked for treatment.

>>And the DoS opportunities that are there today.

Of course and get worse once you can spoof packets to match QoS contracts.

>>How would they provide a usage charge for a set of bits that don't make it through. If that were true, I would ensure a 10% BER on every line to guarantee a LOT of retransmissions and then charge for all of them.

This scheme will be even more effective under the QoS charging plan. But oh wait the target service is VoIP and there will be no re-transmissions!
paolo.franzoi 12/5/2012 | 4:09:21 AM
re: QOS Fees Could Change Everything
unlimited,

You can only buy a different bit rate at the access port. I could give you a GigE interface and put a router with a T-1 connection to the Internet and you would pay for the GigE. You do not pay for a bit rate connection to the Internet (in general and in residential cases). For Enterprises, there are all types of equipment to monitor SLAs.

The reason you are not getting bandwidth measurements is you are paying for the clock rate that the line is wiggling. That's it.

You would only pay for the QoS indirectly. The application provider would pay for the QoS service and you would pay the application provider. Thus, it would be in the best interest of the network provider to provide a quality service that people wanted to subscribe to. This opens up a revenue stream. Without this, there are all types of applications that will never exist.

And the DoS opportunities that are there today. Think about it for a few minutes.

How would they provide a usage charge for a set of bits that don't make it through. If that were true, I would ensure a 10% BER on every line to guarantee a LOT of retransmissions and then charge for all of them.

seven
falsecut 12/5/2012 | 4:09:22 AM
re: QOS Fees Could Change Everything If we must pay for better service, then as a provider, how do I guarantee that my service gets delivered at places that, for example, BST does not control. If the content provider is in Atlanta and agrees with BST to route my traffic through this improved service, how do they make sure that the customer in Denver has the service that the content provider paid extra for. Do they have an agreement with Qwest? If the backbone is provided by a third intermediate, who determines that priority routing?
unlimited 12/5/2012 | 4:09:22 AM
re: QOS Fees Could Change Everything Seven

We already have the option to buy different levels of bandwidth but few if any access providers quantify what they actually deliver. Before we move on to complex QoS schemes (that no one here has so far shown we need) we should get a baseline on what we get for our money today. The utility systems you have shown are at least measurable. Why is it today that we have no idea what we are getting for our money? Why is it not measurable?

Finally of course you don't need to add QoS to implement a usage charge. Access providers are perfectly free to do that today. However, that is determined by the competitive environment. It seems to me that QoS is just a way to have a second swing at pricing. In fact QoS might actually make things worse as it adds more DoS opportunities.
paolo.franzoi 12/5/2012 | 4:09:23 AM
re: QOS Fees Could Change Everything
If one wants to go to the utility model remember:

- Weight of vehicle plays a factor in road taxes (as does gas mileage via gas taxes). The more weight (see tractor trailers) and the lower your mileage, the more you pay to use the roads.

- Sewage and garbage services tend to be billed by the size of the pipe/receptacle.

- Electricity is absolutely pay for use.

The need to charge for QoS is in essence a usage charge. I agree that there needs to be a way to pay for better service. That is what is being proposed. Today's alternative is that there is no way to pay for better performance.

seven
unlimited 12/5/2012 | 4:09:25 AM
re: QOS Fees Could Change Everything "The analogy would be a different cost for a higher-voltage or more reliable electicity supply."

Well it's dangerous to push an analogy too far but since you mentioned it you can pay a one time cost and do this yourself. A transformer can step up the voltage and a UPS and generator can add reliability. This is the difference between a managed or self-provided service. I'm OK with it so long as users have a choice.

On the voice thing. Do we need QoS for voice? In fact what is the reason for putting voice over the Internet? I suggest that it depends on the purpose of the call. Is it time to re-evaluate how we communicate? Given that many people now get cell phone coverage at home (and value the mobility) it is likely that it will replace fixed line telephony service. Perhaps a slight improvement of best-effort will enable good-enough service for those voice applications that can benefit from running over the Internet.
unlimited 12/5/2012 | 4:09:25 AM
re: QOS Fees Could Change Everything It seems that at the root of this discussion is telephony pricing and a need to maintain past revenue performance. That's not an acceptable reason to blockade the Internet which is what is effectively being proposed. With the Internet model it is convenient to have everyone pay to be connected. That fee should get you some simple bandwidth contract. Ideally there would be some fair sharing scheme to prevent bandwidth hogging. You should be able to pay more for a bigger pipe. In this way it's easy for users to put intelligence at the end points to run services and applications. Each network user can implement any local traffic policy for their applications to allocate their bandwidth as desired.

This model fuels innovation and anyone can play. Of course the network can and should be made to be profitable. However, there is no logical argument for why the network providers should get any share of the value in the services or applications run over the network. You would not expect to pay more per watt for the electricity to run a computer than you would for a vacuum cleaner, would you? Now if the network operators want to take part in the services and aplication layers then by all means setup a business unit and compete.

I don't expect this to play out as described because the incumbents have a lot to lose. What I hope is that we get new access competition that will help balance the force. Perhaps there is a role here for municipal wireless networks and even community wireless mesh networks.
PeterU4EA 12/5/2012 | 4:09:25 AM
re: QOS Fees Could Change Everything The notion that all you have to do is 'provide enough bandwidth' to ensure that applications work is, unfotunately, flawed. VoIP, for example, doesn't require that much bandwidth, but it does need low delay and jitter. A single non-differentiated (i.e. best effort) service that meets the needs of VoIP would be massively over-engineered for delay-insensitive applications like file downloads. Whatever the costs of implementing QoS (which needs to go beyond just guaranteeing 'enough' bandwidth), the costs of not implementing it are considerable!
While I agree that there's no logical reason why providers should charge based on the applications they carry, it's not unreasonable they should charge on the basis of the network resources required to carry them. The analogy would be a different cost for a higher-voltage or more reliable electicity supply.
High-Tide 12/5/2012 | 4:09:28 AM
re: QOS Fees Could Change Everything Spelurker is right - throwing bandwidth at a network will never be a 100% reliable mechanism to avoid the need for QoS. Instantaneous bursts of data will occur, and eventually any link will be saturated regardless of size. This argument has raged in IT departments for the last several years. Both QoS and throwing bandwidth at congestion points are workable, sometimes most of the time, but only QoS will work 100% of the time, as long as the network is still alive.

Also, consider the issue of network operation in some sort of a failure mode, when many of the design considerations that were factored into sizing a link may no longer apply. The network needs to be intelligent enough to differentiate high priority traffic under ALL conditions. How would it do so when the link is saturated for some unexpected reason, such as re-directed traffic from another failure, a network storm or DoS event, etc?

Short of instantaneously re-directing (only) priority traffic to a redundant and presumed-healthy link, it seems to me some form of QoS would be required to meet targets of the SLAs.

Admission control is also required, as QoS cannot compensate for insufficient bandwidth when qualified traffic exceeds link capacity.

It's an interesting devlopment that needs to be wrestled with. In regards to the IP-based carrier offerings I've seen locally, their QoS design and the experience level required to advise the customer, and then implement them seem to be lacking. This humble engineer has already found several real-world customer instances where I am concerned about the ability to deliver consistent quality VoIP over these types of IP circuits given the absence of some of the above diciplines.

HT
OldPOTS 12/5/2012 | 4:09:28 AM
re: QOS Fees Could Change Everything Everyone will learn about erlangs inorder to choose their QoS once the QoS comes at a cost, and it will cost.

OldPOTS

I think I've said that many many posts ago.
spelurker 12/5/2012 | 4:09:28 AM
re: QOS Fees Could Change Everything > from every macro analysis, that I've seen, the cheapest way
> to ensure high quality service is to double the connection bandwidth
> of the overloading link. Adding QOS to an existing network,
> can even decrease total throughput.

There is a fallacy with extending this logic though. For it to be workable, the provider needs excess bandwidth at every single link (and node). I've never heard of the being sustainable. All that happens when a link is upgraded is the next link becomes a bottleneck. Especially since, as end-to-end bandwidth availability increases, bandwidth usage always increases to fill it.
This might work for providing point-to-point links, but for providing internet capacity, it falls through. In an internet model, no single provider controls both ends of the connection, and network connection meshes become arbitrarily complex and variable.

The reason people had success with doubling bandwidth is that they were controlling where the bottleneck is. In the case of certain SPs, the bottleneck was the POP. The core was big enough to handle a known amount of traffic based on the size of the incoming links. In the internet case, one of the SP's incoming links is a peering point. This means that your traffic must pass through a bottleneck which you do not control -- the peering point. If your traffic is important to you, it might be nice if there were a way you could ask the SPs to make sure your data passes through the bottle neck.

Now if we could throw enough bandwidth at the problem, then the bottleneck would move all the way into the applications which generate data, and we'd be all done. I don't expect that to happen in my lifetime.
optoslob 12/5/2012 | 4:09:29 AM
re: QOS Fees Could Change Everything QOS is an interesting problem, I know that from every macro analysis, that I've seen, the cheapest way to ensure high quality service is to double the connection bandwidth of the overloading link. Adding QOS to an existing network, can even decrease total throughput. It is akin to paying a BW provider to create an intentional bottleneck and than bidding for priority access to the bottleneck. Just makes no sense at a macro level. ItG«÷s time to dump any BW provider that even suggests this model.

I had hoped the MG would have something to say about walled gardens she has a certain eloquence and moral righteousness that makes the argument compelling. I think my version would sound commercially motivated and crass.

The network manager / provider needs to focus on those technical metrics that make BW cheap, available and useful for an ever larger range of applications.
optoslob






catscan 12/5/2012 | 4:09:30 AM
re: QOS Fees Could Change Everything Take a look at the European RBOCs. They togehter with equipment vendors are standardizing this QoS police. RACS in ETSI Tispan.
I recently read about Tispan at LR,
http://www.lightreading.com/do...
From my understanding the RACS ensures bandwith.

catscan
turing 12/5/2012 | 4:09:30 AM
re: QOS Fees Could Change Everything Spelurker you are so right. Did people think service providers were going to deploy diffserv and MPLS-TE for QoS, and fast reroute and other mechanisms for high reliability, and let some other virtual service provider reap the rewards of that for free?

All this time people have been talking about Qos and such for enterprise VPNs as a good thing, and now they cry when they realize that the service providers are going to offer those enterprises QoS for their service to reach consumers instead of remaining private. (After all, what are Yahoo and Ebay if not big enterprises from BellSouth's perspective?)
TheMuffinMan 12/5/2012 | 4:09:30 AM
re: QOS Fees Could Change Everything Take a look at the European RBOCs. They togehter with equipment vendors are standardizing this QoS police. RACS in ETSI Tispan.
I recently read about Tispan at LR,
http://www.lightreading.com/do...
From my understanding the RACS ensures bandwith.


TISPAN and IMS both have admission and resource control functions - how bandwidth management is done is probably the most interesting part of how to get QoS to work in a converged network, companies like Operax (no I don't work for them) already have RACS compliant boxes, although I don't know of any operator that's yet widely deployed this kind of box.

I think this will be one of the single most important areas for operators to focus on if they want to get out of the bit transport market and find new revenue opportunities!
Arc 12/5/2012 | 4:09:31 AM
re: QOS Fees Could Change Everything The drive for this thread is that G«£Application Service ProvidersG«• (i.e. those that do not own the network) want 90% of the profit margin while assuming 10% of the risk. Now the carriers are starting to fight back. ItG«÷s not surprising they are raising questions regarding net-neutrality. FUD G«Ű Fear Uncertainty and Doubt

It seems that most responding to this article are aware that with out the network there is no service. What is not so obvious is that the carriers have some of the largest capital budgets in the nation and are therefore assuming a substantial part of the risk (bandwidth is not free). It also seems only logical that they should be compensated for providing the infrastructure that delivers the service. The split of the profit margin should be more like 50:50. After all why should long distance transport of IP traffic be free? QOS is one of the mechanisms that will be used to establish a reasonable equilibrium where the network gets paid for. As far as I recall, that is the way a free market system is intended to work.

(Oh ya, regulatory. The only thing that accomplished was the destruction of the worldG«÷s premier R&D lab - Bell Labs. Hmm. I smell sour grapes.)

I have never heard anyone seriously propose that 3rd parties should not have access to QOS. They can buy it if they choose or they can buy Best Effort (BE). If they want low latency and/or more BW then they should pay for it.

WRT to access. I check performance a lot and I have always measured throughput in the access at pretty close to line rate (never use DSL :). So long as I stay on my providerG«÷s network I get very good performance (approaching 15Mbs). As soon as my traffic goes through a peering point ...

For my 2 cents worth G«Ű this transition is long over due.
paolo.franzoi 12/5/2012 | 4:09:31 AM
re: QOS Fees Could Change Everything
I never said I could verify the clocked bit rate of the line. You asked what was advertised and what I agreed to pay for.

When I worked in the Enterprise world there was all types of products with EMS support to ensure that companies were getting the SLAs (it is funny to think of line conditioning as SLAs but whatever) that they paid for.

seven
spelurker 12/5/2012 | 4:09:31 AM
re: QOS Fees Could Change Everything > But if I discover my broadband provider is
> artifically impairing my service (which is
> inevitable even if it is unintentional) or trying
> to make money off of my third-party content
> purchases, I will find another broadband provider.

> In other words, the RBOCs are forgetting who the customer is.

I don't think anyone else is going to make a provider selection based on whether the provider makes money off of 3rd parties. a) No one notices this. b) It makes their services less pricey.

There's companies out there (Skype, 8x8, Vonage...) who are expecting a certain level of service, even though they don't pay for it. I've used Skype -- it's hit or miss. Some calls are pretty good, some are near unintelligible, some are good sound quality, but randomly hang up. These companies have 2 choices -- accept that they will have unpredictable services or they can do something to get reliable service -- they can pay for it.

Why on earth would a bandwidth provider go to the effort of providing differentiated QoS if there was no return on the investment of time and equipment? And even if they already want to provide QoS for one of their own services, they need to perform admission control, to make sure no-one shuts down voip by marking huge FTP transfers as high priority.

The facility-less VoIP providers are living in a bubble, and it's going to burst.
paolo.franzoi 12/5/2012 | 4:09:32 AM
re: QOS Fees Could Change Everything
They advertise the connection bit rate on both cable and DSL (and FiOS for that matter).

seven
unlimited 12/5/2012 | 4:09:32 AM
re: QOS Fees Could Change Everything "They advertise the connection bit rate on both cable and DSL (and FiOS for that matter)."

Have you been able to verify your connection bit rate?
unlimited 12/5/2012 | 4:09:32 AM
re: QOS Fees Could Change Everything "That's the question you should have asked before you signed the initial contract. The answer (if you could've gotten a straight answer) would've been disappointing enough to give you serious pause..."

Not my point. I know what I have paid for but does the average subscriber? If you extrapolate this state of affairs then access providers will charge service providers for QoS but who will verify that it is delivered? Now that could be a good business model!
unlimited 12/5/2012 | 4:09:32 AM
re: QOS Fees Could Change Everything "Actually, what you pay for is a bit rate of the access port. You get no guarantees of anything else. This is massively oversubscribed."

Exactly. But is that what they advertise?
unlimited 12/5/2012 | 4:09:33 AM
re: QOS Fees Could Change Everything "Unfortunately, none of the access providers are going to provide QoS for free. So, your choice of broadband provider is?"

That's a sweeping statement that you need to justify. Also you need to define what you mean by QoS. Then we can start to ask if user's need it.

Here's an interesting question. Today we pay for what is presented as a bandwidth contract with our broadband services. However, we never get close to what access provider's promise. All we get is a theoretical peak bandwidth capability (which is hard to verify) and a huge amount of aggregation into a shared pipe to the actual Internet. It's actually a wonder that VoIP works at all.

Before adding any additional bandwidth management we should clear up what users are actually paying for today.
HeavyDuty 12/5/2012 | 4:09:33 AM
re: QOS Fees Could Change Everything "Before adding any additional bandwidth management we should clear up what users are actually paying for today."

That's the question you should have asked before you signed the initial contract. The answer (if you could've gotten a straight answer) would've been disappointing enough to give you serious pause...
paolo.franzoi 12/5/2012 | 4:09:33 AM
re: QOS Fees Could Change Everything
unlimited,

Actually, what you pay for is a bit rate of the access port. You get no guarantees of anything else. This is massively oversubscribed.

The QoS being discussed here is guaranteeing bandwidth for apps like VoIP or VoD that need better than a best effort service to work all the time.

seven
paolo.franzoi 12/5/2012 | 4:09:34 AM
re: QOS Fees Could Change Everything
Unfortunately, none of the access providers are going to provide QoS for free. So, your choice of broadband provider is?

seven
palaeozoic 12/5/2012 | 4:09:34 AM
re: QOS Fees Could Change Everything
While all of these postings seem to make sense, they miss a fundamental point. I (the residential broadband subscriber) am the customer. I will decide what to do with the service I have purchased. If I want it to go faster, I'll pay more.

But if I discover my broadband provider is artifically impairing my service (which is inevitable even if it is unintentional) or trying to make money off of my third-party content purchases, I will find another broadband provider.

In other words, the RBOCs are forgetting who the customer is.
unlimited 12/5/2012 | 4:09:34 AM
re: QOS Fees Could Change Everything "Converged networks, by their very nature, will provide economies to all users of networks."

This is not intuitively obvious but neither is it obvious that adding QoS is an easy thing to do. QoS will add complexity that will counter some benefits of convergence. Of course it is an attractive idea for Bell South to bill providers but there are many of them and therefore many contracts that will have to be put in place.

Also Bell South can only offer access to their subscribers. How many of the service provider's customers are on Bell South's network and therefore how much should they pay? Will Bell South measure, mark and treat their traffic fairly? This introduces the need for service assurance monitoring to prove service providers are getting what they pay for. And then how should the service providers pass on the cost. Can a VoIP provider charge all its customers the same regardless of whether they get special treatment from their access operator?

If only some access operators implement charging for QoS it may spur more competition for broadband access and offset initial revenue gains. A driver for wireless or powerline access?

How will the scheme work with International service providers and suppose they choose not to pay for their overseas users?

I'm not saying the scheme won't work. Without knowing details it's impossible to say but it will be no slam dunk!
ruready 12/5/2012 | 4:09:35 AM
re: QOS Fees Could Change Everything I agree with these two postings. In ATM and FR you paid for QoS, guaranteed bandwith, etc.

For the service providers to survive, and not be a dumb pipe provider, they need to be more than a bandwidth pipe. QoS, bandwidth guarantees, etc are part of an enhanced service.
krbabu 12/5/2012 | 4:09:36 AM
re: QOS Fees Could Change Everything I agree. Anything of value has to be paid for; and the service providers are justified in charging for QoS through their network. Pure VoIP providers may have a problem with this because of their flawed business models. When they price their VoIP services, they have to build network costs into their pricing. Converged networks, by their very nature, will provide economies to all users of networks.
Enterprise IT departments (that have to provide VoIP service) will build - or lease - high QoS IP networks and can still derive economic use of their IT dollars.
paolo.franzoi 12/5/2012 | 4:09:37 AM
re: QOS Fees Could Change Everything
Selling network capability to get improved service with a business relationship to cover everybody's costs.

Choices that application providers have then are:

1 - Form a relationship, get better service (some sort of SLA will need to be established), and QoS will be part of the arangement.

2 - Use the public internet and get a best effort service. It's cheap, but not guaranteed.

This gives incentive to access providers (whether cable or telco or wireless) to upgrade their infrastructure knowing that popular high bandwidth applications will get them some additional revenue.

seven
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE