In your letter dated Fri, 27 Jan 2012 19:13:46 -0300 you wrote: >> For IPv4, either you have an mtu of 1500 or you use mss clamping. Relying on >> pmtud gives a bad user experience. > >Since IPv4 MTUs can be as low as 296, and since I doubt you clamp the >MSS to such a low value, you still rely on PMTUD. (Hopefully you system >imlements PMTUD + PLPMTUD for blackhole detection).
The whole point of MTU clamping is you set it to whatever is the lowest on paths between you host and the first IX, transit or what not, where the MTU is at least 1500. So if you have 296 locally somewhere, then yes, you would have to set the mss that low. The reason is that too many sending systems are broken. If you don't set the mss, then may up with failing tcp connection in too many cases. >> PMTU blackhole detection is surprisingly uncommon. > >It is implemented in OpenBSD, I recall it bein implemented in Windows... >so I don't think it is uncommon. Maybe there is just too much Linux around then :-) Does implemented in Windows also mean that it is enabled in just about all cases? Otherwise, we're back to square one. >> With random IDs, every packet that arrives has a 1/65536 chance of colliding >> with a packet that is awaiting reassembly. Total chance depends on the >> reassembly timeout and the packets rate. > >And most importantly, the loss rate. Add a bad combination of reassembly >timeout + loss rate + data rate, and you're toast. > >See e.g. <http://www.rfc-editor.org/rfc/rfc4963.txt> There is an interesting calculation there. Effectively, above 20 Mbit/s you are toast anyhow. And let me repeat, this is only an issue for IPv4 links with an mtu smaller than 1280 and where there is also a one-to-one correspondance between IPv6 and IPv4 address in the stateless-NAT box. This gets into the exceedingly rare territory. >> I don't particularly care about the translators. > >There are other sources of this traffic. It has been reported on the >list that this traffic has been found on the wild. I tried to find that message but I can't. Do you have a message ID? -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
