RE: TCP limitations (was Re: ISCSI: Urgent Flag requirementviolates TCP.)

["John H Dowdy" <jdowdy@us.ibm.com>]

Randall Stewart:
>Hmm, seems to me if you implement P-MTU discovery this problem would
>go away, but I do agree that if there is no P-MTU discovery you may
>have a problem with fragments....

In practice, it will be very difficult to attain the design
goals without P-MTU discovery.

I never implemented PMTU Discovery, but I did spend a month working on a
design.
I am not a real fan of it. It is kind of kludgy, and depends on a lot of
things that might
not be there (routers in the path might not send ICMP error packets, and
those that
do might not fill in the MTU). Largely, my dissatisfaction is due to the
probing for
larger MTU which will almost always result in dropped packets and
corresponding
halving of the connection window and drop in throughput.

Can anyone point me to documents that indicate PMTU is all good? I.e.,
bandwidth
can be very high, probing is not necessary, etc.

John Dowdy - SSG Development
Telephone (919)254-5632 (T/L 444-5632)
jdowdy@us.ibm.com