Re: [tcpm] Is this a problem? Fri, 02 November 2007 07:20 UTC

Return-path: <>
Received: from [] ( by with esmtp (Exim 4.43) id 1InqpR-0007Eh-Ac; Fri, 02 Nov 2007 03:20:29 -0400
Received: from tcpm by with local (Exim 4.43) id 1InqpP-0007DN-K0 for; Fri, 02 Nov 2007 03:20:27 -0400
Received: from [] ( by with esmtp (Exim 4.43) id 1InqpP-0007Cl-AH for; Fri, 02 Nov 2007 03:20:27 -0400
Received: from ([]) by with esmtp (Exim 4.43) id 1InqpJ-0003NL-2s for; Fri, 02 Nov 2007 03:20:27 -0400
Received: (qmail 18750 invoked from network); 2 Nov 2007 07:20:05 -0000
Received: from ([]) (envelope-sender <>) by (qmail-ldap-1.03) with AES128-SHA encrypted SMTP for <>; 2 Nov 2007 07:20:05 -0000
In-Reply-To: <>
Subject: Re: [tcpm] Is this a problem?
References: <>
Message-Id: <>
Content-Type: text/plain; charset="US-ASCII"; format="flowed"; delsp="yes"
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v912)
Date: Fri, 02 Nov 2007 00:20:04 -0700
X-Mailer: Apple Mail (2.912)
X-Spam-Score: 0.0 (/)
X-Scan-Signature: 8abaac9e10c826e8252866cbe6766464
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: <>, <>

On Nov 2, 2007, at 12:00 AM, MURALI BASHYAM wrote:

> Clearly there seems to be no consensus on where the problem lies.   
> I've heard that it's not a transport problem, and that the
> responsibility for mitigating the problem lies with a) sockets API,  
> b) OS, and
> c) the application. That's as varied a response as one can get...

There is no need for TCPM to decide which of the three is the best  
solution, since all
three solutions can interoperate on the wire.

The fact that this behavior is not specified by TCP is actually a  
strength. It allows different
hosts to use very different stack strategies. Even for web servers one  
size does not fit all.
What makes sense for a high volume site supporting a large number of  
virtual sites does
not make sense for an embedded web server in a home router that  
supports configuration.

And once you vary the ULP the variations get even greater.

You need to find a more narrowly focused forum that can judge whether  
the proposed
solutions are proper for Web servers (or whatever). But these are all  
issues that the
transport layer needs to avoid.

tcpm mailing list