* Christian Huitema:
The order of5C records in a DNS response is, at best, a
hint. Relying on it as if it were a mandate to clients is a gamble.
When you run RRset-based load balancing, you don't rely on servers
preserving order, or reordering responses. It is completely
sufficient that there is a certain amount of variation among resolver
and application address selection. It has been repeatedly and
independently observed that Rule 9 does not provide sufficient
variance, in contrast to previous behavior.
Rule 9 is also unfortunate because it means that after renumbering,
server loads change in ways the operator cannot influence (except by
requesting addresses with certain bit patterns, but I don't think
anybody wants vanity IP addresses).
--
Florian Weimer <fweimer(_at_)bfk(_dot_)de>
BFK edv-consulting GmbH http://www.bfk.de/
Kriegsstraße 100 tel: +49-721-96201-1
D-76133 Karlsruhe fax: +49-721-96201-99
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf