Subject: Re: Limiting the advertised MSS, again.
To: Rick Byers <email@example.com>
From: None <firstname.lastname@example.org>
Date: 12/07/2001 11:07:19
In message <Pine.NEB.4.33.0112071127590.5815-100000@Apenheul.BigScaryChildren.n
et>Rick Byers writes
>The per-route MTU could easily be used in the calculation of the
>advertised MSS, but I'm not sure its the right thing to do. The per-route
>MTU specifies the maximum size datagram to TRANSMIT on that route. The
>advertises MSS is the maximum size segment you're able to RECEIVE. As far
>as I know, the per-route MTU doesn't effect incoming packets at all.
Tweaking the MSS based on the route MTU is a bit of a lose -- look at
what happens if you send an MSS based on a calculated PMTU, and then
the characteristics of the path change over a long-lived connection.
In that case, you could wind up squeezed into a tiny segment size based
on the smallest PMTU, and you could never probe for any larger MTUs.
In the best of all possible worlds, the MSS would default to 65535 - 40.
>Thats why I was suggesting that maybe we should have per-route MRUs.
This PMTUD blackhole/MSS problem is about the only place where I
can see a use for MRUs. I'd far rather just have the maxmss hack,
instead of all of the extra cruft on each route.
That's just my fairly random opinion. If you're willing to code
it and try it out, don't let me stop you...