> Not to pick on James' post, but several have mentioned that > large routing realms would be one reason to require ULA-G/C > and it has not yet been suggested that the reverse DNS could > itself be thought of as a "large routing realm" of sorts > (depending on how applications use the information they find > there). ULA-G/C would ideally allow collision-free population > of the reverse DNS even with O(10^6) or more delegations.
Not to pick on Fred's post, but making local reverse DNS globally accessible is not recommended. It exposes information which is better kept private. Fred believes that it will make for seamless integration when two organizations decide to merge their local networks. It would probably make the reverse DNS aspect of such a merger more seamless, but it certainly won't make any other aspect any easier. Given that considerable planning will go into all the other aspects of such a merger (securing the virtual connections, bypassing just the right amount of firewalling, provisioning backup links), the effort saved on making reverse DNS work seems negligible. -- George Mitchell > I can't comment [further] on Paul's proposal other than to say > that *some* authority(s) needs to be responsible for the > collision free delegations. I still trust my computer's random number generator more than a centrally-operated entity. -- George Mitchell > Fred > [EMAIL PROTECTED] -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
