ietf
[Top] [All Lists]

Re: way out of the DNS problems? (former Re: delegation mechanism, Re: Trees have one root)

2002-08-01 18:53:35
On 23:17 01/08/02, John C Klensin said:
--On Thursday, 01 August, 2002 22:20 +0200 Bruce Campbell
<bruce(_dot_)campbell(_at_)ripe(_dot_)net> wrote:
> On Thu, 1 Aug 2002, JFC (Jefsey) Morfin wrote:
>> I am sorry to repeat it (and I will try not to say it too
>> many times again): the terms of the ".arpa" sub global
>> namespace delegation are described in RFC 920 by Jon Postel
>> himself.
>
> Request For Comment number 920 is simply a now historic record
> of a policy statement between the IAB and the DARPA in 1984
> regarding establishment of Top Level Domains, and had a focus
> predominately on USA-based organisations.

Bruce perfectly rendered what I said and what is the RFC 920: an historic record of what had been agreed or contracted with "other systems" (cf. RFC 921).

The only interest if that it puts that contracts on the records: WorldCom is not able to find back - up to now - the contract which registered "arapa" with the FCC licensed value added carrier Tymnet. It is true it was for the US namespace but through the FCC licensed IRCs' interconnect agreements in the common frame of the ITU treaties, it was into tInternational namspace.

However, this whole thread is related to the IETF/ICANN position IRT the alt(sic)root. So, to the ICP-3 doctrine which claims its legitimacy from being permanent since a starting point : "The DNS was originally deployed in the mid-1980s" with a link to the RFC 921. RFC 921 describes the DNS implementation according to the RFC 920 rules description.

I have no position on a matter I consider as totally outdated. But I observe that today everything is hampered at the DNSO and at the IETF (in spite of efforts of many of who you are a leading figure) due to a confusion between the DNS system and the public domain it applies to.

So I say:

- either we want to be religious about it like ICP-3, then we must accept that ICANN's lelgtimacy comes from the registration of the ".arpa" domain under standard FCC terms with the rights and constraints reported in RFC 920. This will have some impacts on ccTLD, NSI contract and willl gives different legitimacies to the at(sic)root.

- or (as Bruce calls for it) we drop all this blablabla and we start talking about serious technical issues (my solution 4) through a clean sheet analysis of the technical problems at hand. I only suggest we do not forget all this, to not repeat the same mistake of forgetting about ITU world. As a network we are more concerned by ITU than by WIPO!

I am interested in DNS.2, DNS+ and the service of the users service. Not in old disputes. Am I wrong?

>...
Two additions to Bruce's comments...

(i) While I was not involved in the discussions leading up to, or surrounding, RFC 920, I do have some recollections from earlier and subsequent periods. Given those recollections, and what 920 actually says, I find it fairly difficult to believe that any part of that RFC was motivated by a name-space-division agreement between Jon and/or the ICCB/IAB and a collection of operators of systems based on OSI or other protocols.

Who talks about OSI and protocols? Let have fun at "com/net" coined as a joke in Heindhoven (NL) to be NSI's billions :-)

We talk about the monthly bill of the "arpa" account - under FCC approved rates - which included for free my support of a few British, Argentinean or Japanese people wanting to access "the nets" though their local data service.

But we are also talking of some, puting things in jeopardy on silly pretensions. We are blocked for 2 years for some projects in Europe because ICANN wants to "negotiate" the 100 chars of ".eu" in the master file.

I am very polite and patient, but this is crazy, all the more when you know the joke the use of ISO 3166 2 letters codes comes from...

I hope that Joyce will comment on her recollections when she considers that appropriate.

(ii) As a DNS name allocation policy document, RFC 920 was effectively superceded by RFC 1591.

Respectfully, this is no UDRP. The good faith in managing "arpa" is not questioned. The point is just that a domain name has been rgeistered, not the whole registry purchased. It was a license, not the code..

I was involved in the formulation and writing of 1591, as were several people who I assume follow this list. I am certain that there were no discussions during that process about any commitments to name space constraints or divisions as the result of prior international agreements.

I know... a very good RFC for the network, but which left ccTLDs and Govs without common guidelines. This disinterest is usual with value added customers: you do not commit to any agreement, you respect the way the treaties and regulating bodies have established. But when you grow you happen to share in this.

This is what ICANN tries to sort now. Mike Robert's letter to Govs, GAC involvement, call of Stuart Lynn, ITU increasing involvement, China, EU, today's protestation to the DoC.

But IMHO this is not IETF cup of tea anymore?
jfc


<Prev in Thread] Current Thread [Next in Thread>