On Wed, Sep 16, 2020 at 03:16:45PM +0100, Jeremy Harris wrote: > On 16/09/2020 09:11, Mark Andrews wrote: > > There are a number of issues. > > > Stupid firewalls that block PTB messages. Solution, fragment at network > > MTU. > > Let me rephrase that for you: "Solution, fragment at network MTU (a > value that you cannot efficiently discover)."
i'm not mark but if i were i'd resent you trying to put words in my mouth. in any case please consider whether we are building for today's network or for some larger and more mature future network. if the latter, then RFC 8899 (and RFC 8900) are in scope. this thing went through 22 revisions and came out of that process with its coherence and practicality intact. https://datatracker.ietf.org/doc/rfc8899/ this is why draft-fujiwara-avoid-fragmentation says to use technology like PLPMTUD if it's available, and if it works, use the learned PMTU, and only otherwise, use a prefam MTU estimate. (and there's controversy over both that estimate and the recommendation of setting DONTFRAG on DNS datagrams.) it doesn't have to work today in order to be considered viable and desirable. see also attached. the voiceover and video for it is on the dns-oarc web site. -- Paul Vixie
Fragging DNS.pdf
Description: Adobe PDF document
_______________________________________________ dns-operations mailing list dns-operations@lists.dns-oarc.net https://lists.dns-oarc.net/mailman/listinfo/dns-operations
