[3424] in linux-net channel archive
Re: your mail
daemon@ATHENA.MIT.EDU (Alan Cox)
Sun Jun 23 15:08:33 1996
From: alan@lxorguk.ukuu.org.uk (Alan Cox)
To: dennis@etinc.com (Dennis)
Date: Sun, 23 Jun 1996 17:24:47 +0100 (BST)
Cc: mike@lserv.conexio.co.za, linux-net@vger.rutgers.edu
In-Reply-To: <199606221948.PAA13123@etinc.com> from "Dennis" at Jun 22, 96 03:48:17 pm
> >and random (wind and rain on damaged and poorly insulated local leads out
> >in the bush). It might be able to tell the difference between a *particular*
> >congestion problem and a *particular* kind of flaky line.
>
> Or more often an overloaded router.
You can measure some differences in behaviour but it is hard, and at the end of
the day for most cases the TCP view is that a flaky line is a form of congestion
is good for the net as a whole if not the user. There are experimental things
like SNOOP that sit on a very unreliable end hop (typically radio hop) and
do local retransmission regeneration.
Overloaded routers show some other fun properties like dropping bursts of
frames every 30 seconds or whatever the routing table update schedule for the
router is.
Alan
