On 1/13/16 3:22 PM, Shumon Huque wrote:
On Wed, Jan 13, 2016 at 3:05 PM, Ray Bellis <r...@bellis.me.uk <mailto:r...@bellis.me.uk>> wrote: On 13/01/2016 20:01, John Levine wrote: I suppose, but doing it as _<service>._client._<proto> puts it in > the existing service namespace. It's not a huge difference, and it's > been clear for a while that if we do Dave's registry, part of what > it includes will be a list of the underscore names that are protocol > labels. That currently includes _tcp _udp _sip _xmpp _ldap _http > _ocsp so I suppose adding one more isn't awful, but it seems > needlessly kludgy. When Dave and I last discussed that draft in any detail (back in Orlando!) my proposal was that it should for SRV-based services the only entries should be _tcp and _udp (or other L3 protocols), and that anything that exists in the IANA port registry (with the prepended underscore) would be a legal label to the left of that. That would completely get away from the need to maintain the likes of _sip, etc in the underscore registry. Is Dave's registry proposal documented in written form anywhere? Some planned alignment with that (assuming it has consensus) seems like a good idea. -- Shumon Huque
You mean this? https://datatracker.ietf.org/doc/draft-crocker-dns-attrleaf/ _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop