[ih] forwarded post - need help on history of A-Root server's delegation

[Joe Touch]

(forwarded on behalf of charles.simon at cigref.fr)

PS - some gaps inserted by my rewrapping tool; blame me if they're in 
the wrong place

-- Joe Touch
    Postel Center Director
    USC/ISI

-------------------------------

 > Hi,
 >
 > I'm new here so please forgive me if I break any written or unwritten
 > rules (didn't find any of those 2 on the website anyway). English is
 > also a foreign language to me so pardon my French.
 >
 > I'm currently working at the Cigref which is a French CIO's
 > association (www.cigref.fr). Some of you may know my boss, Sébastien
 > Bachollet, as he is the one reviewing the latest gTLDs'
 > introduction's process for ICANN (fell free to flame on). At the
 > moment, I'm finishing a short paper on the Internet for our members
 > to get a clue on what ICANN and other Internet's actors do and do
 > not. As we were reviewing the latest version of our work, one
 > question rose: who gave NSI/VeriSign the authority to manage the
 > A-Root server?
 >
 > After 3 days working on it, I came up with a little story and I'd need
 > someone to validate it. I Hope some of you could help me out.
 >
 > So here goes the story as I get it based on official documents and
 > some quick e-mails with people from AT&T.
 >
 > In 1998, the Internet Green and White Papers state that "currently,
 > NSI operates the A root server" but it is merely a fact, it does not 
indicate when the delegation first took place. Then the Amendment 11
 > to the Cooperative Agreement Between NSI and the U.S. Government
 > authorized NSI's continued operation of the primary root server. It
 > looks and tastes like a renewal of delegation from the USG to NSI
 > but, then again, it doesn't say when the A Root server's management
 > was delegated in the first place and, most importantly, by whom.
 >
 > I found some writings saying the A root server's delegation to NSI
 > came down from the NSF/NSI Agreement (Cooperative Agreement NO.
 > NCR-9218742) effective from January 1, 1993 on. Someone who was part
 > of InterNIC on behalf of AT&T also thinks it came down from there. He
 > remembers the first day of operation as being 4/1/1993 with
 > presentations later at the IETF plenary session in Columbus. I
 > couldn't check it myself, the Columbus meeting was the 26th meeting
 > of the IETF and the proceedings served on IETF's site start only with
 > the 29th one.
 >
 > What bugged me in this story was that the Cooperative Agreement only
 > aim at registration services and, on October 19, 1992, in the course
 > of discussing its agreement with the NSF, NSI published a so-called 
"Proposal for Network Information Services Manager(s) for NSFNET and 
NREN" which led to InterNIC to be formed (see it there:
 > 
http://www.base.com/gordoni/thoughts/dns-control/rs.internic.net/nsf/nis/proposal-toc.html).
 >
 > In this proposal, it is said (see section K of the proposal): "Network
 > Solutions [...] manages the domain root server for the entire
 > Internet".
 >
 > So I guess they meant the root system, right? So it mean NSI manage
 > the root system before 1993, still right? At the same time, somewhere
 > else in the proposal (Appendix B), it is said NSI was only a
 > subcontractor to Defense Information Systems Agency (DISA) and
 > handled the root system for military uses. After scratching my head
 > quite a bit, I came to the conclusion the meaning of some terms used
 > in the Cooperative Agreement changed in the course of time. Like the
 > term "registration" was pretty much wider at the time that it is
 > today so it mislead me to wrong conclusions when reading the
 > Cooperative Agreement. The result of the Cooperative agreement was
 > indeed to delegate the root's management to NSI directly and for
 > non-military uses when NSI did actually manage it before but under
 > indirect contract with the DISA and for military uses.
 >
> This theory goes along quite well with Article 3.A of the Agreement
("The Awardee shall provide to non-military internet users and
> networks all necessary registration services (which were) previously 
> provided by the Defense Information Systems Agency Network 
> Information Center(the DISA NIC)").
 >
 > That's the story as I got it. Can someone confirms that the facts
 > above are acurate and possibly corrects what's wrong, adds up things
 > I forgot?
 >
 > While working on the subject, I also encountered further questions.
 > Back to the NSF's story, this institution only delegated 3 things in
 > early 1993:
 >
 > (i) Registration Services,
 > (ii) Director and Database Services,
 > and (iii) Information Services for the NSFNET.
 >
> NSI was assigned the first job, AT&T the second, General Atomics the
 > third. Following NSI's proposal, the 3 delegees joined up to build up
> InterNIC. In December 1994, when the Midterm Evaluation by NSF took
 > place, General Atomics got dismissed. Then, in december 1997, NSI got
> InterNIC out of the IP allocation's loop (Amendment 7 to Cooperative 
> Agreement Between NSI and U.S. Government) and ARIN showed up. From 
> then on and until ICANN popped up, I don't know what happened.
 >
 > So, I just have 2 questions on that subject, maybe some old-timers
 > could answer them:
 >
 > - does someone know where to find the original agreements between
 > AT&T and the NSF (Cooperative Agreement NCR-9218179), General Atomics
 > and the NSF (Cooperative Agreement NCR-9218749)?
 >
 > - how did AT&T and NSI get along regarding InterNIC until InterNIC
 > was passed over to ICANN and who did what (respective jobs of NSI,
 > AT&T)?
 >
 > I heard AT&T trademarked InterNIC at the time and then passed it to
 > the DoC. If it's true and as I read somewhere else NSI tried to fly
 > away with InterNIC's name and homepage when ICANN jumped in, I guess
 > it should have been ugly at the time. Overall, Internet history is
 > quite a mess to me so I hope you could help me getting my ideas
 > straight on those few points.
 >
 > Thanks for reading this so far, I hope I'm not spamming, I sent a
 > previous version last week but it seems it didn't make it to the list
 >
 > Charles
 >
 > Ps: oh, also, I know root-servers' operators are generally volunteers
 > picked up by Postel at the time but my questions are really specific
 > as it seems the process for delegating the A-root is kinda special.
 > Thanks again.