John C Klensin wrote:
... There is a strong case to be made that the introduction of the underscore convention was a kludge that violated fundamental design assumptions of the DNS and that it was added without considering, much less acting on, what other changes would be needed to support it smoothly. ...
jon postel raised the same point (as the rfc editor when SRV was published.) i'll tell you what i told him: we needed only one thing, which was an identifier that could never conflict with a "host name". to do that, we added a character (_) to the front of these service and transport names which was not in the syntax of the old HOSTS.TXT definition.
it was minimal, and not intended to be generalizable, but it violated _no_ design assumption, fundamental or otherwise, of the DNS.
where SRV was a process violation is that it tried to cover existing systems like the then-young "world wide web" with some load balancing logic but without any geo-ip logic which has since been shown to be widely desired.
however, SRV works for the people and protocols who use it, and the experience we gained from "names that pass in the night" underscoring has informed the community's long term understanding of what we need.
jon didn't love it but he understood it and he withdrew his objection. i support creating a real registry for future reserved-word DNS labels. i just don't want to apply it to SRV. -- P Vixie _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
