Re: [aqm] 答复: [iccrg] [tcpm] ECN support and usage on the Internet

["Scheffenegger, Richard" <[e-mail]>]

Hi Simon,


For one, throughput is the one metric, that should NOT be influenced by ECN, actually. So if that is "in the noise", that is a very good sign; 

The more interesting metrics, that are addressed with ECN are latency ("page load times") and network efficiency (fewer packets are needlessly processed in initial hops, just to be dropped by the penultimate hop; that bandwidth can be allocated to other flows etc).

Other schemes that reduce latency usually do so by trading a good share of bandwidth... (credit allocation scheme used in FC, anyone?)


ECN hat a huge issue of deployability issues, that had nothing to do with the mechanism itself, but with misbehaving equipment on the signaling - and the institutional memory still remembers those albeit any new signaling scheme could run into similar issues (independent of the mechanisms).


Ie. some people still haven't got the Memo that IPv4 TOS was deprecated quite some time ago (now DiffServ/ECN),

Regards,

Richard Scheffenegger



> -----Original Message-----
> From: [e-mail] [mailto:[e-mail]] On Behalf Of
> Simon Barber
> Sent: Mittwoch, 20. März 2013 00:43
> To: Jim Gettys
> Cc: Rick Jones; [e-mail]; [e-mail]; 天澜; [e-mail]; Mikael
> Abrahamsson
> Subject: Re: [aqm] 答复: [iccrg] [tcpm] ECN support and usage on the
> Internet
> 
> I'm not sure why people are so obsessed with ECN - the little data I've
> seen about it showed that it's effect on system throughput was in the
> noise. Sure, wireless guys don't like to waste precious wireless
> bandwidth, but they're not dropping packets that have used up the precious
> wireless bandwidth. They get dropped before the wireless link.
> 
> Simon
>