On Thu, Jan 22, 2015 at 12:06:44PM -0500, Tim Wicinski wrote: > > The two options are alternative proposals - there should never be a reality > where both proposals exist as supported options.
Right. The question is whether you think people are going to behave well (which is more or less what Ray's proposal enables; the server if under attack just closes its TCP anyway), or whether you prefer a mechanism where the server gives hints that guide a client's behaviour. I prefer the latter approach (which is what's in Paul's and Joe's draft). I think it provides data by which a server can develop profiles of those requesters that reach it, and thereby make more intelligent decisions about resource allocation. But that would be complicated and tricky, and Ray's proposal has the really notable advantage that it's quite simple. So I'd like one of these documents to go ahead. A -- Andrew Sullivan a...@anvilwalrusden.com _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
