[76812] in North American Network Operators' Group
Re: Smallest Transit MTU
daemon@ATHENA.MIT.EDU (Dan Hollis)
Thu Dec 30 14:31:05 2004
Date: Thu, 30 Dec 2004 11:29:58 -0800 (PST)
From: Dan Hollis <goemon@anime.net>
To: Florian Weimer <fw@deneb.enyo.de>
Cc: Jerry Pasker <info@n-connect.net>, <nanog@merit.edu>
In-Reply-To: <873bxoggmv.fsf@deneb.enyo.de>
Errors-To: owner-nanog-outgoing@merit.edu
On Thu, 30 Dec 2004, Florian Weimer wrote:
> * Dan Hollis:
> > Because tcp connection endpoints have to implement ECN in order to manage
> > the flow.
> Your wording suggests that ECN is purely an end-to-end signaling
> protocol
it does? where?
> (and so does a lot of propaganda from the ECN zealots).
an "ecn zealot" is someone who wants firewalls to work correctly? someone
who wants idiots to stop blocking all icmp a "pmtud zealot"?
> But is this really true? If I read the RFC correctly, you need *routers*
> that use ECN to indicate congestion instead of packet drops.
anything along the path can *indicate* congestion, but its up to the
*endpoints* to *respond* to the ECN indication and mitigate their flows.
read rfc3168 paying close attention to 6.1.2 and 6.1.3
-Dan
