- From: Patrick McManus <mcmanus@ducksong.com>
- Date: Tue, 15 Dec 2015 14:31:05 -0500
- To: Phillip Hallam-Baker <phill@hallambaker.com>
- Cc: "ietf-http-wg@w3.org Group" <ietf-http-wg@w3.org>
Received on Tuesday, 15 December 2015 19:31:32 UTC
if you think of a srv as a more fine grained cname, the answer is obvious - don't change the host header as it needs to reflect the origin. On Tue, Dec 15, 2015 at 12:29 PM, Phillip Hallam-Baker < phill@hallambaker.com> wrote: > I am trying to spec out a Web Service. It is obviously desirable to use > SRV based discovery over A-record since that allows for priority, fallback, > etc. > > However rfc7230 is silent on the matter. > > The point of complication that enters is what the name for the Host: > header. Should this be remapped to the address given in the SRV or not? My > view is that it should not. > > so if the service being resolved is _mmm._tcp.example.com and we have srv > records: > > _mmm._tcp.example.com SRV 0 5 80 host1.example.com > _mmm._tcp.example.com SRV 0 5 80 host2.example.com > > The Host: header should be: > > Get /.well-known/mmm/ > Host: example.com > > > This seems like the obvious approach. However there are two other > possibilities: > > Host: host1.example.com > Host: _mmm._tcp.example.com > > The first is obviously unhelpful and wrong as the SRV alias identifies a > machine, not the site. > > The second could actually be useful though. It essentially eliminates the > need for .well-known URIs. > >
Received on Tuesday, 15 December 2015 19:31:32 UTC