[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gpsd-dev
Subject:    Re: [Gpsd-dev] Cosmic Background Radiation Bufferbloat Detector
From:       Ed W <lists () wildgooses ! com>
Date:       2011-04-14 14:38:21
Message-ID: 4DA706DD.6080207 () wildgooses ! com
[Download RAW message or body]

On 12/04/2011 03:48, Eric Raymond wrote:
> One of the things I work on other than GPSD is the Bufferbloat project
> aiming at solving Internet perfornce problems caused by over-buffering and
> poor queue management.  Dave Taht of that project is currently my houseguest
> here at World Domination Central and we've been brainstorming ways to
> write bufferbloat diagnostic tools.

Surely if you can't detect the problem using straightforward TCP
timestamps then the problem is too subtle to care about?

As I understand the "bufferbloat" claim, it's about routers queuing an
excessive number of packets and so network latency might remain constant
right up to the point the speed exceeds bandwidth available, at which
point it increases dramatically due to queuing?

Measuring this seems more like a network stack level curiosity and in
any case a transitory event?

I should have thought there was more chance to get big name ISPs to turn
on ECN (hence getting it well debugged across the internet) and then
buffering becomes a non issue?

Ed W
_______________________________________________
Gpsd-dev mailing list
Gpsd-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/gpsd-dev
[prev in list] [next in list] [prev in thread] [next in thread]