Re: [tcpm] TCP zero window timeout?

Joe Touch <touch@ISI.EDU> Sun, 27 August 2006 02:59 UTC

Received: from [] ( by with esmtp (Exim 4.43) id 1GHAsJ-00005J-5D; Sat, 26 Aug 2006 22:59:51 -0400
Received: from [] ( by with esmtp (Exim 4.43) id 1GHAsI-00005E-KV for; Sat, 26 Aug 2006 22:59:50 -0400
Received: from ([]) by with esmtp (Exim 4.43) id 1GHAsH-0004hw-9h for; Sat, 26 Aug 2006 22:59:50 -0400
Received: from [] ( []) by (8.13.6/8.13.6) with ESMTP id k7R2whJ1003993; Sat, 26 Aug 2006 19:58:43 -0700 (PDT)
Message-ID: <>
Date: Sat, 26 Aug 2006 19:58:43 -0700
From: Joe Touch <touch@ISI.EDU>
User-Agent: Thunderbird (Windows/20060719)
MIME-Version: 1.0
Subject: Re: [tcpm] TCP zero window timeout?
References: <>
In-Reply-To: <>
X-ISI-4-43-8-MailScanner: Found to be clean
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 7aafa0432175920a4b3e118e16c5cb64
Cc: "Mahdavi, Jamshid" <>,, "Anantha Ramaiah (ananth)" <>, Fernando Gont <>
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: TCP Maintenance and Minor Extensions Working Group <>
List-Unsubscribe: <>, <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
Content-Type: multipart/mixed; boundary="===============1685694265=="

> Fernando
> I collaborated with mahesh on this, so let me try
> making the case for it a little better.
> The problem was found on a TCP proxy, which does not
> have any applicaton awareness. In fact application
> awareness is not a goal in that environment. Hence the
> TCP level solution.

Application-aware proxies are one thing (and they would be able to
adjust fine here). TCP proxies for splice-like, transport-only gateways
are already known to break so many things it's not clear that
considering that as a special case is a good argument for modifying TCP.
It's a better argument not to use TCP that way.

> Having said that, i'd say this timeout is in the same
> spirit as the upper bound on the retransmit mechanism
> of TCP. TCP could indefinitely retransmit and have the
> application timeout too, correct? The problem is that
> TCP has a persist state which can potentially exist
> infinitely long  and which lends itself to abuse by a
> malicious peer. 

TCP is not intended to be robust to security attacks. A peer that can
establish a TCP connection is already presumed to be a non-attacker at
the TCP level, IMO.


tcpm mailing list