> > The whois database is part and parcel, a necessary element of DNS
> > operation. It is impossible (or at least unreasonable) to conceive of
> > running a TLD zone file without keeping track of who is associated with
> > each second level domain.
>
> Actually, that's incorrect. The DNS SOA record shows who produced the
> record and according to the RFC's the postmaster@... address is mandatory.
> There is just no technical *NEED* to have more data.
SOA records in a zone file are not relevant to the contact database.
I'm not talking about data that is published, but rather the data needed
to operate a TLD class zone file -- if one is going to perform billing
(per Amendment #4) or simply validate changes or expire require renewal
after a period of time, it is necessary to have a contact database right
alongside the zone file itself.
A zone file without a contact database is like a a phone company that
retains no records of who to bill for each number. It is not a matter
whether that such a database is made public or not, only that it must
(and, in fact does) exist.
> In fact .NA does not give out this information for its registrants and
> that has been working for almost 8 years.
But I would bet that there is an database of who the registrants are, what
domain each has, when the registration expires, etc. It is this database
that I am talking about.
> > The two parts - zone file and whois/contact database have always been
> > two parts of a whole. The whois/contact database is useful in and of
> > itself, but the zone file starts to disintegrate and becomes rapidly
> > worthless without the whois/contact database -- without the contact
> > database it is impossible to validate updates, or even to collect fees.
>
> That's opinion, not fact. You actually strengthen NSI's argument, if you
> look at it closely: They were hired to produce the ZONES without which the
> Internet doesn't work.
No, they were hired to perform registrations. Registrations produce two
things -- 1) a record of who performed what registration and 2) a zone
file.
The former is the contact database.
> They ALSO produced, the WHOIS data base. As far as I am concerned, this is
> their intellectual property and they have all the rights in the world to
> market it
They have to produce a whois equivalent database just to do the job which
is required of them by the cooperative agreement.
They may not have to publish it (although the Internic group of contracts
requires that they do), but they certainly need it to perform their
internal billing and renew operations.
> > Indeed, the fact that NSF approved fees contains an implicit statement
> > that there is a contact database upon which the contractor, NSI, can
> > administer to process renewals.
>
> Yes, and? Does it say this must be handed over?
Yes, the cooperative agreement says very explicitly that NSI must turn
over (if the government requests) all data necessary for the government to
enable a sucessor to pick up the registration job for the TLDs run by NSI.
And for that to happen, not only must the zone files be conveyed, but also
the database of contacts and expirations dates.
> > The whois database is about as ancillary to DNS as wings are ancillary to
> > a flying airplane.
>
> Nonsense!
How do you propose to perform billing without it? How to you propose to
know whether someone is sending in an authentic update without it?
The contact database is absolutely part and parcel of the job of running a
TLD registry on the scale that NSI is doing.
And, moreover, it was the standard way of doing things *before* NSI
started its job under the contract. The Internic contract is
extraordinarily vague about deliverables and can only be interpreted in
the context of the way things were done before by SRI and BBN.
As such, the US government has the contractual right to obtain a copy of
the whois database, in addition to the raw zone files, at the end of the
cooperative agreement.
--karl--
