On Fri, May 31, 2013 at 2:02 PM, Ted Hardie <ted.ietf@gmail.com> wrote: > On Fri, May 31, 2013 at 1:44 PM, Eric Rescorla <ekr@rtfm.com> wrote: > I personally think S-NAPTR could work here, but there are two trade-offs > to consider: > > S-NAPTR targets are domain names, which means you may incur a round-trip > in discovering the associated addresses for the domain names. In some > cases, they are pointers to services which also require S-NAPTR look ups > (see section 4.3 of RFC 3958 for an example); those may incur two round > trips. There are ways to optimize this, but it can incur extra latency. > > S-NAPTR (and other DDDS approaches) really give you two things: data > about what's available and data about the ordering that is preferred by the > service owner. While that might come into play here, I have heard a use > case for it. Is there one > And in this case, that data is in JS that came from the server, so it could express those opinions directly in the API. -Ekr
Received on Friday, 31 May 2013 21:06:43 UTC