ARIN LIVES!!!

[Gordon Cook]

June 24: The National Science Foundation has just announced the formation
of ARIN, (The American Registry for Internet Numbers). NSF has approved a
Network Solutions plan to set up the independent IP registry. All
necessary documents are signed and in place. Steps to establish ARIN begin
immediately. The enabling breakthrough came in negotiations between the
parties last week in Washington.

>From the July - August COOK Report on Internet - published today.

Administration Approves Formation of ARIN, pp. 1-10

In an eleventh hour decision last week, the Clinton Administration dropped
its opposition to an American Registry for Internet Numbers. Ira Magaziner
finally grasped what was at stake and, to his credit, acted forcefully to
end foot dragging by other agencies.  We applaud these events.  For the
approval for NSF to set up ARIN actually takes the first important and
coherent step towards the Administration's announced goal of industry
driven self- regulation.

Nevertheless, the road to last week's decision was marked by a remarkable
amount of bungling and lack of both coordination and leadership among
federal agencies that, in the feelings of some, were more interested in
protecting their turf and in not making "wrong" decisions than in really
trying to understand the key reasons behind the crisis in Internet
governance. Several issues compounded the problem. First, OMB, because of
its role in coordinating implementation of the Federal Administrative
Procedures Act, took a major role in deciding what should be done.
Unfortunately the players at OMB had no understanding of the complicated
historical, political and legal linkages that they were dealing with when,
having been called on to fix Domain Name Service, they decided also to
meddle in IP numbers.

Second, while the policy makers felt there were issues of control over
business critical elements of DNS as well as uncertainties about the role
of the ITU and viability of the IAHC process, the Inter Agency Task Force
set up to deal with the issue had no leadership worthy of the name. Many
of the stake holders had no grasp of the technical complexity and legal
linkages between what undoubtedly first appeared to them as nothing more
complicated than intellectual property aspects of obtaining business
addresses in Cyberspace. One person directly involved expressed dismay to
us at the enormous gulf he saw between the concerns of the network
engineers and the non technical policy wonks - something that he simply
did not understand but intuited to be serious.

Finally, in such a context, the only way to avoid disaster, was for those
interested to lobby the top policy makers such as Magaziner and do
whatever it took to educate them. We report with considerable relief that
this seems to be what was finally accomplished last week. What we have
seen however is only Act One. There is still much that remains to be
decided about DNS policy. Some court cases are underway that will likely
force rapid decisions. Some are also asserting that the "old boys network"
of Internet governance is dead and that the commercial Internet industry
must now throw out the consensual processes that have been the foundation
of the Internet's growth and prosperity.   These people we have little
respect for. Therefore, in order to spread awareness of what has happened,
we present a detailed summary of the behind the scenes maneuvering of the
past two months.

In April ARIN was back on track and headed for a September 1 opening,
when, suddenly at the beginning of May, we received word that ARIN was
once again on hold. Why? Because OMB had decided to fix the problems of
IP. The only problem was that the underlying problems, which are
technical, are not administratively "fixableÓ, and the people sitting
around the Inter Agency DNS Task Force table either didn't know it or
would not admit it. After sending scathing private mail to an
administration official, we received a reply on May 11 that told us worlds
about the problem. "As far as I know -- the only outstanding objection to
ARIN is whether they are dealing with number portability.?  Certainly --
number portability is critical in the telephony context to promoting
competition  - so people are asking -- why not portability for Internet?
If you have any recommendations for people on the technical side - I'd
appreciate it."

We passed this data along to the appropriate technical leadership of the
net, went to Russia and waited for more news. When it came it was that a
succession of technical folk had done the educating called for but that
amazingly ARIN had been thrown a new curve. The feds were now insisting it
be announced in the Federal Registry before it was formed. We were told
that this new delay would kill ARIN, and that worse, it was doing nothing
to solve the authority problems of the IANA.

On June 2 Ken Cukier published an extremely important article in
Communications Week International. It detailed the Dublin meeting of RIPE
the European Registry that had occurred a few days earlier. There it was
announced that both RIPE and APNIC had made monetary contributions to take
up the slack in IANA funding in view of ARPA's non renewal of the contract
with ISI that had paid for many of Jon Postel's functions. For the more
knowledgeable here was a clear implication that, if the US government
didn't do something to stabilize IANA's problems, Postel could simply move
this critical piece of Internet governance outside of the US. We sent the
article to our electronic subscribers as an "extra" and later heard that
the contents had found their way to the June 3rd meeting of the
Interagency DNS Task Force where they had had a significant impact in
raising the level of consciousness of what was at stake.

In the meantime, on the way home, a visit to London (June 9 -13) enabled
us to discuss matters with Tony Rutkowski, some key US regulators, and the
CIX President and Executive Director. These meetings gave us a more
balanced view of the position of the "other side". The CIX in particular
indicated a feeling that ARIN was not urgent and that it was time for
commercial providers working through a US Government "process" to reshape
the way the net was run. Discussion with others after our return from
Russia and London has left us viewing the CIX position with considerable
dismay.

With ARIN on hold, needed fundamental changes in IP policy could not be
made. But with dues of $1000 a year, we'd estimate that 4 out of every 5
ISPs in the ARIN service area could afford to join. Since the member ISPs
will elect the ARIN Policy Council, it seems to us that on sheer numbers
alone, ARIN can be an effective mechanism against unwarranted industry
consolidation.

We have had serious doubts about the Clinton Administration's Internet
policy. In this instance the Administration did the "right thing." We hope
that it will continue to do so on heels of challenges that will follow.

USISPA Lawyer Fails to Understand Needs of ISPs, pp. 11 - 13

On April 28, in some critical discussions on the NAIPR mail list, USISPA
lawyer Rudolph Geist painted ARIN as device hatched by a monopoly in order
to become another monopoly and perpetuate the interests of Network
Solutions and the big service providers. To the response of Philip Nesser
"I don't believe that the process should be completely open to the public
(the finances yes, but not technical applications) because the information
requested may be considered proprietary by many organizations;"
Geist replied: "It is highly suspicious to maintain that technical
information (or any  information for that matter) regarding the allocation
of IP address  blocks, a finite public resource (like telephone numbers or
radio  spectrum), should be held proprietary by a monopoly outgrowth
(ARIN) of  another monopoly (Internic)." It developed that at least in one
instance his way of trying to get an address block for his clients was by
threatening to sue InterNic or force it to break the procedures set forth
in RFC 2050. We believe that neither Mr. Geist nor his organization
deserves respect.

Seven New gTLDs Make No Business Sense, p. 14

A short cogent critique by Vanderbilt Professor Donna Hoffman

UUNET and Sprint Charging for Peering, pp. 15 - 23

At the very end of April, in an explosion that rocked the Well and its ISP
Whole Earth Networks (WeNet), David Holub, WeNet's founder refused an
order from Well owner Bruce Katz to capitulate to UUNET's demand for paid
peering. Holub was fired for taking what we view as a courageous stand
against UUNET's insidious policy (since modified) of insisting that anyone
who wished to even talk about remaining a UUNET peer would have to sign a
non disclosure agreement that would presumably keep the fact that paid
peering was under discussion and the price paid for peering a secret.
UUNET actually had bought about three months of silence with this policy
before Holub's courageous action blew the whistle. 

In the meantime Sprint began a policy by which its peers would have to pay
in declining amounts, according to the number of exchange points where
they conducted peering. Sprint's prices represent an average cost increase
of $200,000 a year for each peer. If we look at the big five privately
peered providers, we can see that, should all of them start doing what
Sprint and UUNET have done, the average increase in operating expenses for
the peers would be a million dollars a year per backbone. While Sprint and
UUNET offered technical justifications for their charges, the cost benefit
assessment of who should settle with whom in a connectionless network is
still no more clear than it has been in the many many discussions of these
kinds of issues that have cropped up on NANOG and elsewhere over the last
couple of years.

What does seem obvious is that there may be strong anti-competitive
elements lurking in these new policies. Roughly two years ago
the big five established criteria for free peering. Since then between 15
and 20 national high speed backbones have appeared that meet the criteria
- far more than many ever imagined possible. Now that this has happened,
at least some of the big five are changing the rules in such a way as to
put a financial squeeze on their own would be competitors. We present our
own analysis of these events and a summary of mail list discussion from
early May through mid June.

Holub's Model of Open Peering Open Interconnect Internet, pp. 24 - 26

David Holub went to the California PUC and managed to have WeNet declared
a common carrier. In a NANOG discussion with Kent England, he presents his
strategy for open peering.

Choosing an Upstream Provider and the Cost Equation for Dial Up Service,
pp. 27-29, 48

On inet-access Sean Donelan provides valuable advice on the variable to
consider when shopping for backbone service. Also welcome news is ANS's
recent serious entry into providing backbone service for ISPs. In a second
discussion, several ISPs respond to Avi Freedman's query on the costs of
providing their dial up service.

NSF Inspector General's Defunct Plan to Tax Domain Names, pp. 30 -35

This absurd plan was officially rejected on April 17. We publish our
analysis of it and the plan itself in full as a monument to the kind of
thinking that real policy makers should in future avoid.

State of Russian Internet, p. 36

Brief reflections on internet developments in Russia.

Book Reviews, pp. 37 - 40

Reviews of recent O'Reilly books on Java and the Web by Russian
programmers in St Petersburg.

K - 12 Technology Debate, pp. 41 - 46

Part 2 of a debate between Ferdi Serim and Jeff Michka.   Part 1 appeared
in the Feb 97 COOK Report.


************************************************************************
The COOK Report on Internet               For subsc. pricing & more than
431 Greenway Ave, Ewing, NJ 08618 USA     ten megabytes of free material
(609) 882-2572 (phone & fax)              visit   http://cookreport.com/
Internet: cook at cookreport.com             On line speech of critics under
attack by Ewing NJ School Board, go to http://cookreport.com/sboard.shtml
************************************************************************