[Iccrg] RE: Iccrg Digest, Vol 44, Issue 6

[Ingemar Johansson S]

Hi

I am a bit curious regarding the interperetation of the word supported
(or not supported).

To me "ECN not-supported in the routers (or whatever)" can mean two
things

1) The ECN bits are ignored and left unchanged
2) The ECN bits are zeroed.

The first alternative is not serious for the endpoint, also the opinion
I have seen sofar is that core routers does not want the extra workload
of peeking into the inner details of the IP headers (may not even look
at the IP headers at all...). The principle as I understand it is that
the edge nodes are supposed to experience congestion.

The second alternative is of course more serious for the endpoints that
decide to use ECN.

So my question is what kind of behavior (1 or 2) is to be expected from
e.g core routers that don't support ECN  ?

Regards
/Ingemar

> Message: 2
> Date: Wed, 29 Jul 2009 13:30:12 +0200 (CEST)
> From: Mikael Abrahamsson <swmike at swm.pp.se>
> Subject: Re: [Iccrg] Is ECN too complicated?
> To: Bob Briscoe <rbriscoe at jungle.bt.co.uk>
> Cc: iccrg IRTF list <iccrg at cs.ucl.ac.uk>
> Message-ID: <alpine.DEB.1.10.0907291324200.10769 at uplift.swm.pp.se>
> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
> 
> On Wed, 29 Jul 2009, Bob Briscoe wrote:
> 
> > If ECN is still too complex, I'm at a loss (irony intended).
> 
> Well, let's put it this way:
> 
> The major core routing platforms today do not as far as I 
> know support ECN at all, and these are the ones with big 
> buffers (and support gazillions of queues etc). I asked for 
> ECN support and received pushback for reason being "I was the 
> only one who asked for this" (basically).
> 
> So ECN support is present in the CPU-based routers from a 
> certain major vendor, but not the more ASIC/NP-accelerated ones.
> 
> So as I stated in the meeting, the ISP community is going 
> more in the direction of "let's make a 100GE router/L3-switch 
> with minimal buffers that never congests and upgrade to" than 
> "let's find this platform that can squeeze the max out of 
> this 10GE we already have".
> 
> DWDM-capacity is cheap, doing buffering/intelligent things in 
> routers is expensive.
> 
> I can go more into detail about this if someone feels it's 
> appropriate.
> 
> -- 
> Mikael Abrahamsson    email: swmike at swm.pp.se
> 
> 
>