LET'S JUST GO AROUND

[Howard C. Berkowitz]

Karl,

Up front, I'd like to thank you for some of your comments.  I may disagree
fervently with you at times, but just as often, you make me think.

At 11:46 AM -0800 2/5/97, Karl Auerbach wrote:
>> It would *definitely* be 'restraint of trade' if you can't get a /19
>> from the NIC until you're "big enough"...
>
>It's my feeling that ARIN itself would be pretty safe from this sort of
>complaint if they:
>
>	- Have a well articulated policy about this
>
>	- Express that policy clearly to applicants for address space
>	  *before* fees are paid and let the applicants know the risks
>	  of getting an address that some ISPs won't route.
>
>	- Have a well expressed and applied conflict-of-interest policy
>	  for their relationships with ISPs.  (As usual, I'm not
>	  expressing this very well... what I'm trying to get at is that
>	  there are a lot of potential ISP relationships with ARIN due
>	  to BoT and advisory council memberships, and we need to prevent
>	  even the appearance that ARIN is, in-effect, an ISP-owned body.)\

This is hard to express also.  At some level, I am not as worried as I
might be about this.  Now, I am the first to admit that the Internet is not
in the same environment as the telecom carriers.  Yet there are parallels.

AT&T, when it was Ma Bell, took on certain functions by industry consensus.
These included North American Numbering Plan admistration, the telco part
of the national emergency Restoration Priority System, etc.  After Judge
Greene spoke, some of these functions went to Bellcore, while others went
into other industry forums.  I am drawing a blank on the operational one,
as it's been a while ... Inter-Carrier <something mumble> Forum.  There was
also the Exchange Carriers Standards Association, and, as ISDN became
common, the North American ISDN Users Forum (which had a significant
carrier piece).

The IC-whatever-forum had some aspects of NANOG, but was a much more formal
activity.  NANOG serves a very useful function and is not broken.  I do
wonder, however, if there is getting to be a need for more of an
inter-carrier policy organization dealing with the broad sense of ISP
issues, routing policy, issues.

ARIN is not the place for this sort of discussion.  IETF or ISOC
conceivably could be, but have not stepped up to the role.  Perhaps no one
can.  Perhaps the partially regulated telecom industry is too different.  I
don't know, but I don't want to keep diverting the ARIN discussions into
out-of-scope policy issues.

Two obvious problems arise in creating an organization to deal with
intercarrier policy:  the appearance of collusion for restraint of trade,
and the not-unrelated perception of lack of user input.  Smaller ISPs have
aspects of both users and carriers, and don't fit neatly into the model.

My experiences in dealing with the first problem date from my 5+ years at
the Corporation for Open Systems (COS), which, for those of you that don't
remember, was a not-for-profit consortium intended to accelerate the
adoption of interoperable OSI and ISDN products.  Lots of things done wrong
there, but some things are relevant.

COS was a membership organization with a substantial staff -- up to about
140 when we were doing active development of protocol test systems.  I was
employee #4, and the first technical person on the staff.  While most of my
time was in software development, product certification,and education, I
still spent time in secretariat functions.

When the organization was first set up, the Justice Department was invited
in to review matters to avoid antitrust.  Certainly for the first couple of
years, until the DoJ was satisfied we did not pose an antitrust threat, the
lawyers were very involved, to a ridiculous extent at times.  But one of
the secretariat jobs was to watch for suggestions of prohibited collusion,
and intervene IMMEDIATELY.  For example, if during a technical meeting, a
member got up and uttered even a couple of words about pricing or his
company strategy, in those terms, I was REQUIRED to jump up, YELL If
necessary, and warn him to cease immediately.  There were cases where the
staff had to go to the mat with members and stop meetings.  But it could be
done.

We never established a satisfactory way to get user input, which in large
part came down to the reality that most users did not want to be involved
in the continuing process and expense.  They expected their vendors to do
it.  Certain large users, such as the Defense Department, General Motors,
etc. did take an active role, typically representing an industry segment as
well as their own organization.
>
>As for the ISPs that block -- well, I think that they may eventually have
>a lot of explaining to do.  The reason for this is that the incremental
>cost of carring a route for a /24 is the same as for a /8.  If the ISPs
>want to be "common carriers" (and hence obtain many protections against
>being liable for the content of the traffic they carry) they may have to
>fairly offer their services to all comers.

Now you've got me struggling to articulate something.  At one level, you
are quite correct to say the incremental cost of carrying a /24 is the same
as carrying a /8.  Yet, at a different level, the cost of carrying the set
of /24 prefixes is much greater than the cost of carrying the set of /8
prefixes.  There is demonstrably more economic and technical risk into
agreeing to the principle of carrying the set of /24 prefixes -- and we
probably can't do it.  So incremental cost isn't quite the term we want,
although it is a factor.  Perhaps someone can come up with a better term,
but to me incremental cost has to be considered in the context of the size
of the theoretical number of increaments -- of quanta -- in the universe of
discussion.
>
>		--karl--