Return-Path: <jhawk@MIT.EDU>
Received: from central-city-carrier-station.mit.edu by po11.mit.edu (8.9.2/4.7) id QAA24978; Fri, 21 Sep 2001 16:21:01 -0400 (EDT)
Received: from manawatu-mail-centre.mit.edu (MANAWATU-MAIL-CENTRE.MIT.EDU [18.7.21.85])
	by central-city-carrier-station.mit.edu (8.9.2/8.9.2) with ESMTP id QAA21321;
	Fri, 21 Sep 2001 16:20:59 -0400 (EDT)
Received: from multics.mit.edu (MULTICS.MIT.EDU [18.187.1.73])
	by manawatu-mail-centre.mit.edu (8.9.2/8.9.2) with ESMTP id QAA10245;
	Fri, 21 Sep 2001 16:20:59 -0400 (EDT)
Received: (from jhawk@localhost) by multics.mit.edu (8.9.3)
	id QAA00551; Fri, 21 Sep 2001 16:20:58 -0400 (EDT)
Date: Fri, 21 Sep 2001 16:20:58 -0400
From: John Hawkinson <jhawk@MIT.EDU>
To: "McNerney, Tim" <TMcNerney@chipwrights.com>
Cc: "'sipb@mit.edu'" <sipb@MIT.EDU>
Subject: Re: Who do I talk to about e19-rtr-backbone.mit.edu lossage?
Message-ID: <20010921162058.L3186@multics.mit.edu>
References: <93A4BBB9515ED411AE9D00508BD955A5198C1F@intranet>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <93A4BBB9515ED411AE9D00508BD955A5198C1F@intranet>
User-Agent: Mutt/1.3.19i

McNerney, Tim <TMcNerney@chipwrights.com> wrote on Fri, 21 Sep 2001
at 16:06:53 -0400 in <93A4BBB9515ED411AE9D00508BD955A5198C1F@intranet>:


> The basic problem I am trying to address is that low-bandwidth, 
> interactive TCP sessions (e.g. ssh/telnet, X) are getting 

Most people do *NOT* consider X to be low-bandwidth.

> particularly screwed when routers drop packets on the floor, 
> which, of course, is the "acceptable" way of dealing with network 
> congestion.  

I'm not sure what you're suggesting should be going on here. There's really
nothing else you can do in a congestion situation than drop packets. That's
basically what congestion is.


> It feels to me that the fundamental problem is that while network
> bandwidth is increasing, it is no longer being allocated in a
> way that is "fair" to interactive sessions.

You say, "no longer," as if previously there was some sort of "enforced
fairness." I think that is not accurate.

You are, of course, free to suggest to MIT's network group that they
try to consider some sort of prioritization method that prioritizes
some kinds of traffic over other kinds. This is generally considered a
non-trivial problem.

I don't mean to suggest that SIPB would be unable to work in an advocacy
role for you here, but I do think that this is not a problem that
is common to a great many people. Of course, that characteristic 
may make it easier to solve.


> The problem may be getting worse as routers get optimized for
> "bursty" traffic.

I do wonder what you mean here. It sounds like you are misinformed,
but perhaps you could provide more detail.

> Right now the MIT backbone accommodates lots of different types of
> traffic:  big packets, small packets, images, MP3s, streaming video, 
> etc.  The extremes are file transfers that need throughput but can 
> sacrifice latency, and interactive sessions that really need low 
> latency but only constitute a trickle's worth of bandwidth.
> 
> What I had in mind was having two (real or virtual) networks inside
> MIT, one low-bandwidth, low-latency, and one for everything else.
> The point being to make it so that interactive streams don't get 
> "starved" the way they are now.  
> 
> Without doing lots of research, I'm not sure I flesh this out too much 
> more, but I hope you get the idea.  

Ah. You are not proposing a seperate physical network. See above.

>    [jhawk continues]
>    Off the cuff, it seems like a pretty niche service.
> 
> Don't people still use ssh and remote X servers?  That's the sort of
> "niche" I am talking about.

I thought you were proposing a seperate physical connection to MIT.

I think the set of people who ssh into MIT is smaller than you might suspect,
but perhaps not.

--jhawk
