Re: [tcpm] tcpsecure: how strong to recommend?

Joe Touch <touch@ISI.EDU> Sat, 29 September 2007 21:21 UTC

Return-path: <>
Received: from [] ( by with esmtp (Exim 4.43) id 1Ibjkf-0006Fc-1J; Sat, 29 Sep 2007 17:21:29 -0400
Received: from tcpm by with local (Exim 4.43) id 1Ibjkd-00068J-Bg for; Sat, 29 Sep 2007 17:21:27 -0400
Received: from [] ( by with esmtp (Exim 4.43) id 1Ibjkd-00068B-2A for; Sat, 29 Sep 2007 17:21:27 -0400
Received: from ([]) by with esmtp (Exim 4.43) id 1IbjkW-00053J-PW for; Sat, 29 Sep 2007 17:21:27 -0400
Received: from [] ( []) by (8.13.8/8.13.8) with ESMTP id l8TLKkS6011555; Sat, 29 Sep 2007 14:20:46 -0700 (PDT)
Message-ID: <>
Date: Sat, 29 Sep 2007 14:20:37 -0700
From: Joe Touch <touch@ISI.EDU>
User-Agent: Thunderbird (Windows/20070728)
MIME-Version: 1.0
To: "Mitesh Dalal (mdalal)" <>
Subject: Re: [tcpm] tcpsecure: how strong to recommend?
References: <>
In-Reply-To: <>
X-Enigmail-Version: 0.95.3
X-ISI-4-43-8-MailScanner: Found to be clean
X-Spam-Score: 0.9 (/)
X-Scan-Signature: e1b0e72ff1bbd457ceef31828f216a86
Cc:, "Anantha Ramaiah \(ananth\)" <>,
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="===============0363869794=="

Mitesh Dalal (mdalal) wrote:
> FWIW, I think we need to address the fundamentals first, as echoed by
> Lars, Tim and your truly
> some time back. i.e. are we labeling this changes as updates to rfc793 ?
> Will we clearly state
> (if the document proceeds to  standards track) to state "Updates: 793"
> in the title of the document ?
> OR
> are we looking it as an additional protocol that can be added/embedded
> to TCP (although there is no handshake to signal this understanding
> between the TCP peers to distinctly picture this as an option,
> but a small protocol nevertheless) and does not necessarily update the
> base spec ? 

Whether this is a small change or not is irrelevant to whether it
updates TCP's rules for handling segments. All modifications to TCP
segment processing events update 793 - whether they are standards track
or experimental.

An "additional protocol" could not change any TCP headers, TCP endpoint
state, *or* TCP segment handling events.

I remain confused at this sort of question coming up at all. If this is
not a change to TCP, at the very least it is being discussed in the
wrong working group.


>> -----Original Message-----
>> From: Mark Allman [] 
>> Sent: Friday, September 28, 2007 5:24 PM
>> To: Anantha Ramaiah (ananth)
>> Cc:
>> Subject: Re: [tcpm] tcpsecure: how strong to recommend? 
>>> In the last meeting in Chicago, it was pointed out the only pending 
>>> issue is the "strength of the mitigations". 3 choices were listed, 
>>> people picked the choices ( Careful not to use the "vote" since you 
>>> don't seem to like it :-), So I would think the next step 
>> is to pick 
>>> on whichever choice comes up as popular and move on. I for 
>> one don't 
>>> see anything wrong with that approach.
>>> Atleast I am missing as to what would constitute a "rough 
>> consensus" ?
>> First, it isn't me that doesn't like the word "vote".  It is 
>> the IETF that doesn't take "votes".  Taking "votes" is 
>> problematic.  Take a look at some of the IETF process 
>> documents, the Tao of the IETF, etc.  These explain why there 
>> are not votes within the IETF and the rationale behind the 
>> consensus process.
>> It seems to me that in this case it is clear that the WG is 
>> not generally of one mind (i.e., come to consensus), with a 
>> non-trivial number of folks wanting MAYs, SHOULDs and and/or 
>> some combination (via picking some MAYs and some SHOULDs or 
>> using conditionals or whatever).
>> If you can show us that we're reading this wrong and that in 
>> fact these folks are outliers then please do.
>> allman
> _______________________________________________
> tcpm mailing list

tcpm mailing list