Return-Path: <ghudson@MIT.EDU>
Received: from pacific-carrier-annex.mit.edu by po11.mit.edu (8.9.2/4.7) id NAA29480; Fri, 28 Sep 2001 13:40:28 -0400 (EDT)
Received: from egyptian-gods.MIT.EDU (EGYPTIAN-GODS.MIT.EDU [18.101.0.162])
	by pacific-carrier-annex.mit.edu (8.9.2/8.9.2) with ESMTP id NAA03436
	for <sipb@MIT.EDU>; Fri, 28 Sep 2001 13:40:26 -0400 (EDT)
Received: (from ghudson@localhost) by egyptian-gods.MIT.EDU (8.9.3)
	id NAA14093; Fri, 28 Sep 2001 13:40:24 -0400
Message-Id: <200109281740.NAA14093@egyptian-gods.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?
In-Reply-To: Your message of "Fri, 28 Sep 2001 12:03:26 EDT."
             <93A4BBB9515ED411AE9D00508BD955A5198C31@intranet> 
Date: Fri, 28 Sep 2001 13:40:24 -0400
From: Greg Hudson <ghudson@MIT.EDU>

> at least a solution that allows me to keep from losing my ssh/telnet
> session several times a day.

This may not be a problem with MIT's network.  A number of friends of
mine have run into problems with firewalls or NAT gateways (on their
end) which can lose track of TCP state and reset the connection the
next time they receive a packet.  Characteristics of this lossage are:

  * You lose your connection instantly after typing or otherwise
    causing a packet to be sent, usually after a period of idle time.

  * Your sshd process on the MIT side is still around and is still
    waiting for data on the network socket.

Of course, the second characteristic can also mean a network failure
reported to the ssh client but not to the ssh server.

> perhaps TCP (and its lame retransmit policy) is no longer "the right
> thing."

A more aggressive retransmit policy could degrade network performance
for other users, so calling the TCP retransmit policy "lame" is
probably prejudicial.

> People have retorted that no one cares about telnet/ssh anymore, but
> I think it is a sad state of affairs if "the net" has become a
> "ground freight only" medium at the expense of "air-mail letters."
> (i.e. lightweight, high-priority packets).

I wouldn't say the web is "ground freight."
