Re: [pmtud] Improvement for the current PMTUD mechanism

At 08:08 a.m. 17/10/2005, John Heffner wrote:

> > Actually, this cannot have interoperation problems. IIRC, TCP congestion
> > control is not mandatory.
> > And even then, fast retransmit is not specifically congestion control. If
> > anything, I'm proposing to slow down the packet rate, rather than
> > increasing it. (i.e. being more conservative).
>
>In point of fact TCP congestion control is a MUST per RFC 1122, but this is
>entirely separate from the issue at hand. :-)
>
>Any change to standards constraining how or when TCP may recover from a loss
>is pretty significant.  I think it is a discussion topic for tcpm.

After reading the PLPMTUD draft, I think this is kind of sooting yourself 
in your own foot.

My solution to the scenario you raised was to disable fast-restransmit when 
there's a pending ICMP message.

Disabling fast-retransmit is okay, as it is being more conservative (you 
send *less* data).

PLPMTUD disables *congestion* *control* which leads to more data being 
sent, and could lead to congestion in the Internet.

Yes, it may not be a killer, but it's certainly much less conservative than 
disabling fast-retransmit.

Kindest regards,

--
Fernando Gont
e-mail: fernando@gont.com.ar || fgont@acm.org

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