& cplSiteName &
Comments
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
<<   <   Page 8 / 12   >   >>
sgan201
sgan201
12/5/2012 | 1:49:49 AM
re: Siemens Sees Ethernet Everywhere
Hi,
The goal of networking is to deliver bits from point A to point B and in the process support the applications that the user want to support. On the access/edge network, if the bandwidth is cheap, you could sacrify bandwidth and use equipment that do not support strigent QOS. If the bandwidth is expensive, use better QOS equipment to achieve efficiency. There is no silver bullet and one size fit all solution out there. 10 years ago, the BS was ATM everywhere and even ATM to the desktop. My response was it does not make sense to ATM on LAN since Ethernet was a better solution. So, now, the BS is Ethernet everywhere.. Now, my response again is let use some common sense here.

1) If you have fiber to a location and bandwidth is free to you, use the cheapest equipment..

2) However, if your bandwidth is Oc-12 and below and you have to support more than IP and/or you need to control jitter and provdie absolute QOS guarantee, ATM sound like a better choice..

3) If you are less than DS1/T1/E1, FR is a better solution..

4) If you are less than 64K, run X.25..

There is no one size fit all out there..

Dreamer
stephenpcooke
stephenpcooke
12/5/2012 | 1:49:49 AM
re: Siemens Sees Ethernet Everywhere
In case anyone was wondering about the protection switching protocol that would need to be used for those G.SHDSL collector rings that I mentioned...

Don't even think of UPSR (Unidirectional Path Switched Rings), they are far too bandwidth inefficient. The only choice is BLSR (Bidirectional Line Switched Rings) with Extra Traffic capability. This gives full bandwidth availability in both directions for non-essential traffic.
Tony Li
Tony Li
12/5/2012 | 1:49:48 AM
re: Siemens Sees Ethernet Everywhere

I'm sure that the cable indeed has some fairness mechanisms built into it. However, it is certainly on Ethernet in my house and I believe that it's on Ethernet & POS within my provider's network.

Are you saying that QoS at one point in the network is enough?

Tony
Tony Li
Tony Li
12/5/2012 | 1:49:47 AM
re: Siemens Sees Ethernet Everywhere

Dreamer,

I agree completely. Let's use some common sense. The word "Ethernet" no longer denotes anything other than an L2 frame format. You can have QoS on an Ethernet just like any other point to point media. Just don't use an under-buffered switch that doesn't recognize your QoS identification and there is no issue.

Tony
stephenpcooke
stephenpcooke
12/5/2012 | 1:49:46 AM
re: Siemens Sees Ethernet Everywhere
In a collector ring where all traffic are aggregated into a hub, there is no difference in bandwidth efficiency between UPSR and BLSR. BLSR advantage comes in only when your traffic pattern is distributed.

In the case described, UPSR transmits the same information, originating at any node, in BOTH directions around the ring. For example: If your bandwidth around the ring is 10Mb/s in each direction (ie: your hub has a total incoming bandwidth of 20 MB/s) and you generate 1 MB/s of priority traffic at each of your 10 nodes, UPSR is 100% filled. BLSR is 100% filled in ONE direction with the other direction completely available for extra traffic which gets dropped in the case of a line cut.
sgan201
sgan201
12/5/2012 | 1:49:45 AM
re: Siemens Sees Ethernet Everywhere
Hi Tony,
QOS mechanism is needed when there is possibility of congestion affected the supported application. So, depending on target applications and offerred load and so on, you have to design the network accordingly. And, the access and edge network has to be designed separately due to the constraint of the bandwidth is usually the highest at that point..

What is right for the core usually is not the right answer for edge. Network design is an art and science that should not be degenerate into a slogan of "ATM everywhere" or "Ethernet everywhere".

We could advance network design skill and knowledge if people just not stop getting religious and qualify their argument like assuming there is fiber to every location, this "X" is a better for this location.. Assuming there is only IP traffic on this network, "Y" might be a better solution so on..

Dreamer
sgan201
sgan201
12/5/2012 | 1:49:44 AM
re: Siemens Sees Ethernet Everywhere
Hi Stephenpcooke,
Are you sure that you are not comparing a 2 Fiber UPSR versus a 4 Fiber BLSR??

Dreamer
sgan201
sgan201
12/5/2012 | 1:49:43 AM
re: Siemens Sees Ethernet Everywhere
Hi,
I had both cable modem service and DSL lines. I work from home and I have VoIP. Cable modem is no matched compare to DSL line in reliability and consistency of performance.. It shows up especially when you use a VoIP service like Packet8.

My friend did online gaming via IDSL (ISDN DSL) -> 128Kbps.. When he switched to cable modem, he can no longer play game consistently eventhough the bandwidth supposed to be higher.. Now, he is switching back to DSL..

Dreamer
sgan201
sgan201
12/5/2012 | 1:49:43 AM
re: Siemens Sees Ethernet Everywhere
Hi,
1) Everyone is on a separate VCC/connection. ATM policing can make sure that fairness is provided. A bandwidth hog will not kill everyone else IP connection. Everyone get a guarantee amount of bandwidth even during congestion. This is one of reason why VoIP service like Vonage generally works better at DSL line as opposed to cable modem. Cable modem's folks had implemented DOCSIS 1.1/2.0 to provide some of this. But, this is not part of IP..

Yes, you can do this via IP too. But, then, you have use a add on box to provide traffic shaping/policing and so on. This will work on TCP/UDP layer. It does not support as many connection as ATM and it is expensive.

2) ATM chopping of packets into cells allow better control of jitter. The last link to the CO is usually DS1 IMA/OC3/OC-12.. Better control of jitter give you lower delay variance.. If you do a ping for DSL system versus cable modem, yuo will see that the round trip time for DSL has lower variances between min and max time..

3) ATM switches coming from WAN legacy are designed with huge buffer that allow it to handle network congestion without dropping IP packet -> less retransmission -> better throughput..

4) Per VCC queueing/buffering mechanism in the ATM switch allow for during congestion time only drop IP packet from the bandwidth hog that over burst their tolerance without punishing anyone..

5) ATM OAM Loop back capability allow the service provider to do a ATM loop back test on you DSL modem. And, they can do this in regular interval without involving IP layer. They can fix and know the problem before you call them. Supporting large number of users without truck roll is the goal here.

6) Most of connection between ATM switch are done through SPVC. ATM can reroute the connection and respect the QOS guarantee and the same time before IP router layer know this. Less link down -> less re-transmission -> better performance for user.

ATM connection bandwidth is guaranteed even during network congestion and re-routing.

The bottomline is very simple. People will judge for themselves. Do they prefer cable modem or DSL line if they have a choice?? Will they prefer DSL line with ATM versus DSL line with IP DSLAM.. We will all know eventually..

"The truth is out there"

Dreamer
Flower
Flower
12/5/2012 | 1:49:43 AM
re: Siemens Sees Ethernet Everywhere
Dreamer, you keep telling us that we need ATM to get good QoS. Please explain, as I can't see how ATM is gonna help my IP traffic.

Everybody uses IP to transport their data. Agreed ? People at home and people at the office will have ethernet NICs in their PCs. Agreed ? Therefor data will leave PCs in IP over ethernet.

Now somewhere in the network, those IP packets will be transported inside ATM. If I understand correctly, ATM can only give different QoS to different VirtualCircuits. So my IP packets have to get multiplexed into different VCs to get QoS. How is the device that is gonna translate IP/ethernet into IP/ATM going to know which packets need to go into which VCs ? That device needs to know about all IP protocols, applications and portnumbers ? Does my PC need to set up new VCs for different destinations/applications/requirements ?

I have ADSL at home. My IP packets travel over ATM (Nobody told me when I got my subscription, but I actually was able to notice and measure the overhead involved with ATM !). But I get only one VC from my DSL router/modem to the DSLAM. How is ATM going to give me QoS here ?

IMHO the only way that IP traffic can get QoS is when the applications mark IP traffic with certain priorities (obviously via the TOS or Precedence bits). If IP packets are marked this way, IP routers can do the prioritization themselves. Nothing ATM can add.

So I really would like to understand how ATM can give me (the user) QoS that IP can't. TIA.
<<   <   Page 8 / 12   >   >>


Featured Video
Upcoming Live Events
October 22, 2019, Los Angeles, CA
November 5, 2019, London, England
November 7, 2019, London, UK
November 14, 2019, Maritim Hotel, Berlin
December 3-5, 2019, Vienna, Austria
December 3, 2019, New York, New York
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events