Re: [Ipsec] Issue 91 - ICMP error message handling

[Mark Duffy <mduffy@quarrytech.com>]

Hi Barb,

Karen has since re-posted her ICMP proposal of April 13, so I suppose it is 
that one that should be on the table, rather than the older one from Oct 26 
2003.

Neither of those proposals covers the path MTU discovery.  So I think PMTUD 
needs to be carried forward from 2401 (modified or not) in any 
event.  Whether to act on Karen's proposal should be a separate 
issue.  (Your item #3 seems to me to suggest that Karen's proposal is an 
*alternative* to the PMTUD text.)

--Regards, Mark


>In the original 2401 there was text describing the handling of path MTU
>but the handling of other unprotected ICMP messages was left as a local
>matter.  On October 26, Karen posted the message included below that
>proposes a more complex algorithm for handling ICMP messages.  There has
>been little discussion to date on this so Ted and I would like to
>propose the following:
>
>1. Default position: if no one is interested in discussing this issue,
>we fall back to the text as it existed in 2401 which covers path MTU but
>is silent on other types of unprotected ICMP messages.
>
>2. If you believe there are deficiencies or ambiguities in the original
>text on path MTU in 2401 please speak up!
>
>3. If you feel we should take a different approach, such as that
>proposed by Karen, then start voicing your opinions.
>
>We plan to give the working group 1 week to begin discussion on this
>topic. If there is no discussion within 1 week, we will go with the
>default position.  Otherwise, if discussion commences, we're willing to
>invest a couple of weeks to reach a consensus direction on this issue.


_______________________________________________
Ipsec mailing list
Ipsec@ietf.org
https://www1.ietf.org/mailman/listinfo/ipsec