Hi Diego, On Mon, Nov 12, 2012 at 02:13:00PM +0000, Diego R. Lopez wrote: > Hi Sebastian, > > On 9 Nov 2012, at 22:00 , Sebastian Kiesel wrote: > > I think we could even do that within the ALTO protocol, see > > http://tools.ietf.org/id/draft-kiesel-alto-alto4alto-00.txt but I also > > believe that in a multi-domain scenario this inter-ALTO-server mesh > > structure would be much more cumbersome than relying on DNS, and a DNS > > resolver stub is anyway included in virtually every networked device. > > > Was not aware about this draft, thanks for the pointer. I think this could > be the base for what I mentioned in my last reply to Michael on relying > on only-ALTO documents to make the specification. I still think that a > single GET query sounds simpler than a direct access to the DNS interface, > as well as the trust considerations are applicable. > > What would you think of leaving the current discovery document as it is > now and exploring the possibility of aligning HTTPS discovery with this > ALTO4ALTO idea?
As you can see the ALTO4ALTO draft has expired quite some time ago. We've been discussing architectural options for quite a while now and there was not much positive feedback about this draft so far. My personal favorite topic is still the third-party discovery. If we had a working solution for that we could also use it for the resource consumer initiated discovery. But 3pdisc causes lots of headaches so we decided to push forward a simple solution at least for resource consumer initiated discovery. From a technical point of view I think both is feasible, DHCP/DNS based discovery as well as ALTO4ALTO style discovery, but I cannot see a clear winner. From procedural point of view the DHCP/DNS way has advantages as it is based on more established and solid base protocols. Thanks Sebastian _______________________________________________ alto mailing list alto@ietf.org https://www.ietf.org/mailman/listinfo/alto