On Wed, 24 Dec 2003 [EMAIL PROTECTED] wrote: [...] > I have no problem with a SHOULD for stub resolvers. While most > don't do it there is no real unknowns in saying that they should > do it. The caching server they are using most probably is already > making EDNS queries on their behalf.
While DNS servers seem to have added support for EDNS0, the stub resolvers don't seem to be supporting at least widely (e.g., at least BSD, Linux and Win2K don't). At least looking around with tcpdump, I see *very* little UDPsize options in DNS packets, probably coming from newer DNS servers out there which have toggled on the option in their config. EDNS0 is probably a nice thing to have, but as far as I can see, generic nodes don't seem to be doing it now (which is what the doc talks about). I guess the question is about whether we want to encourage adding that support, or documenting existing, well-known practice for new implementations. If we put EDNS0 as a SHOULD, we'd probably be doing the former .. which is OK by me as long as we're doing a conscious decision on that (but my personal take is that we should stick to "known good, implemented, works" -mantra in a document like this). -- Pekka Savola "You each name yourselves king, yet the Netcore Oy kingdom bleeds." Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings -------------------------------------------------------------------- IETF IPv6 working group mailing list [EMAIL PROTECTED] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
