To:
"Eric Brunner-Williams in Portland Maine" <brunner@nic-naa.net>
Cc:
<ietf-provreg@cafax.se>, <brunner@nic-naa.net>
From:
"James Seng/Personal" <James@Seng.cc>
Date:
Fri, 26 Jan 2001 06:54:49 +0800
Sender:
owner-ietf-provreg@cafax.se
Subject:
Re: Merging RRP and Whois
> If you would cite or somehow motivate a requirement for a registry > protocol to provide access to 3rd-parties, regardless of the modality > of the operation (anonymous read vs write, etc.) it would assist my > understanding. RRP is not just a registry protocol. It does not prevent a registrar to use it with their reseller. On allowing 3rd party to access the RRP servers, a) DNSSEC. Keys needs to be exchanged directly with the Registry. b) Charter TLDs may requires direct communication with registrant. I could also cite examples outside DNS space but I will pass for now. > Which of the registrant, registrar, registry interests are augmented > by access via the provreg protocol of parties who are neither registrant, > registrar, nor registry? Some clearly are depricated, either by anonymous > 3rd-party access, or by the (possible) delay in the standards process, > even if experimental, by mission creep in the specifications phase. Disagree. We are not here to fight for anyone interest. We are here to design a standard. And such a standard protocol have implication beyond the 3 groups of people you stated here. > > And IMHO, there is no technical reason to forbid a registrant to access > > the registry directly > > Please share your solution to the aaa problem, on one hand I've on the order > of 100 or 1,000 or so endpoints a registry, or any set of registries, must > authenticate, etc., and on the other hand I (or rather _you_) have got on > the order of 10,000,000 endpoints a registry, or ..., must authenticate, etc. 100 or 1,000 or 10,000,000 or 1,000,000,000,000. If we design it correctly, we can do it. One often hear that "IETF only have one problem: Scalability". This is no exception. It is not easy but running away from scalablity is not an option. > As neither necessity nor sufficiency tests appear to get any more clear > than what looks like "don't know", the proponents of the claim haven't > made a case. Don't know is a valid answer considering that there are factors which we are not able to answer yet. And until we do know these factors, like what what protocol and what infrastructure is been proposed, my best answer is 'don't know'. > > Great. I look forward to the day we repeat this all over again in some > > other WG formed to solve this "out of band" registration problem. > > Iteration with selective scope is an alternative to single pass with unlimited > scope. No. Iteration with selective scope is a typical software design methology, not a design methology. We dont come here to design version 1 only to release version 2, version 3 etc in future. Things dont happen this way in IETF. -James Seng