To avoid further confusion on who said ... <snip> <snip> <snip> ...
The last message was from Jaap Akkerhuis, who said: > > Oops, apparently Alfred said so. But who sais what is irrelevat on the > discussion. The oint I was making is that there should not be a fixed > aministrative model. > > jaap However, I didn't suggest that either. I noted that there is an additional logical entity to consider (i.e., if you properly unsplice the roles), painted an extension to Antoin's picture to show where it "most likely" lives in his entity map, and then continued with pointing out that there are alternatives for the direct contractual relationships. In fact, I see a parallel to the "SIP trapezoid": The direct (RTP) media path between the communication peers corresponds to DNS transactions between the child and parent DNS operators, and the SIP signalling path through <n> proxies and intermediaries corresponds to the chain of organizatorial entities and relationships in our picture. In both cases, the 'signalling' transactions carried out in the 'upper arc' control what eventually happens in the 'bottom arc'. In both cases, what most matters eventually is the timing in the conceptional 'bottom arc'. The possible diversity in the upper arc seems to be an argument in favor of a solution employing direct communication in the bottom arc. It _may_ be that studying this parallelism (and the experiences with the SIP trapezoid) in more detail could result in useful insight(s) -- but please don't overstress this comparison! Kind regards, Alfred. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
