Re: Strange Attractors in Network Flows

Valdis.Kletnieks@vt.edu Mon, 29 July 1996 11:23 UTC

Received: from ietf.org by ietf.org id aa20730; 29 Jul 96 7:23 EDT
Received: from cnri by ietf.org id aa20726; 29 Jul 96 7:23 EDT
Received: from murtoa.cs.mu.OZ.AU by CNRI.Reston.VA.US id aa00376; 29 Jul 96 7:23 EDT
Received: from mailing-list by murtoa.cs.mu.OZ.AU (8.6.9/1.0) id SAA29604; Mon, 29 Jul 1996 18:53:51 +1000
Received: from munnari.OZ.AU by murtoa.cs.mu.OZ.AU (8.6.9/1.0) with SMTP id SAA29553; Mon, 29 Jul 1996 18:33:42 +1000
Received: from black-ice.cc.vt.edu by munnari.OZ.AU with SMTP (5.83--+1.3.1+0.56) id IA09723; Mon, 29 Jul 1996 18:33:39 +1000 (from valdis@black-ice.cc.vt.edu)
Received: from localhost (valdis@LOCALHOST [127.0.0.1]) by black-ice.cc.vt.edu (8.7.5/8.7.3) with ESMTP id EAA34486; Mon, 29 Jul 1996 04:33:28 -0400
Message-Id: <199607290833.EAA34486@black-ice.cc.vt.edu>
To: Tim Bass <bass@linux.silkroad.com>
Cc: big-internet@munnari.oz.au
Subject: Re: Strange Attractors in Network Flows
In-Reply-To: Your message of "Mon, 29 Jul 1996 03:47:45 EDT." <199607290747.DAA32085@linux.silkroad.com>
Sender: ietf-archive-request@ietf.org
From: Valdis.Kletnieks@vt.edu
Pgp-Action: signclearpem; rfc822=off; originator="<Valdis.Kletnieks@vt.edu>"
X-Url: http://black-ice.cc.vt.edu/~valdis/
References: <199607290747.DAA32085@linux.silkroad.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Id: <26798.838629207.1@black-ice.cc.vt.edu>
Date: Mon, 29 Jul 1996 04:33:28 -0400
Precedence: bulk

On Mon, 29 Jul 1996 03:47:45 EDT, Tim Bass said:
> The 'ethernet capture effect' discussion you attached isolates
> one particular effect that in weakly causal in IP internetworks
> and is not central to IP-CHAOS model development and strange
> attractors in dissipative systems.

So what you're saying is "interesting, but not good enough?"

> IP-CHAOS theory would look at the seeming random packet arrival
> and departure rates (or flows rates) and attempt to look for

Well, a chaotic system will only look chaotic if you look at the
right phase space.  Are you sure that it's packet arrival/departure
that is interesting, or flow rates?  Personally, I think a better
(and more productive) line of attack would be looking at routing
flaps (which often exhibit "ringing" as one after another router
falls over while attempting to swallow a large routing update),
or maybe some other measure.  Another thing to look at might be
access rates to popular mirror-FTP sites (especially those that
have made attempts to setup a frobozz.org DNS that tries to round-robin
effectively).  What damping/oscillation rate do we end up seeing
due to the "this FTP server is full" scenario?

Also, are you sure that you would know chaotic behaviour if you 
saw it?  And more to the point, are you likely to *find* it before
some network admin spots it and stomps it out?  I know around here,
we have a tendency to go find and "fix" any sudden oscillation in
packet rates.  I wouldn't be surprised if most of the major long-haul
providers are similar - this may mean that your best shot at finding
chaotic behavior *is* on a small network (a la the ethernet capture).


/Valdis