At-Large Representative for Canada and the United States on the
ICANN Board of Directors
Karl Auerbach

Diary - January 21, 2002

There was a telephone meeting of the ICANN board on January 21, 2002.  As usual it was held at the crack-of-dawn California time (6am).  I had been up late writing code the night before so 6am came far too early for my liking.

As usual the meeting was closed to the public yet was attended by various people who are not board members, who had no particular purpose for being there, contributed nothing of value, and whose presence served only to impair the candor of discussions among the Directors.

The preliminary report may be found at: http://www.icann.org/minutes/prelim-report-21jan02.htm

As has become usual for ICANN board meetings, the agenda is established with minimal advance notice, thus making it very difficult to read, much less, comprehend and be prepared to make informed decisions.  For that reason I had to abstain from some matters.  I can only presume that the other members of the board had found the time to read and and were able to make well-informed decisions on these matters and were not simply being blindly led by "staff" recommendations.

Reconsideration Requests

ICANN's "reconsideration committee" chanted out its usual anthem that "ICANN can do no wrong".  I am more than suspicious that these "reconsideration" reports are written by "staff" and not by the committee itself, a fact which, if true, would raise a serious question in my mind whether the reconsideration committee, taking a cue from the Board as a whole, has simply abrogated its responsibilities to ICANN's management.

In view of the nearly automatic rejection of  just about every reconsideration petition ever filed,  I find it disheartening that there is still no visible progress towards the Independent Review Panel.  I guess that the IRP will arrive arm-in-arm with Godot and the Chimera.  In other words, I am doubting that ICANN will ever actually establish the Independent Review Panel that it has been obligated to create for several years now.

As for the matters at hand, I abstained on the reconsideration reports RC-01-4 and RC 01-7:  I did not feel adequately prepared to make a decision on these.

As for RC 01-6:  ICANN's staff apparently lost this request for the better part of a year.  I voted to accept this recommendation mainly because I felt that the particular issue complained of was handled.  However, when examined in the broader context of UDRP decisions, I believe that ICANN must begin to seriously consider the degree of competence, or lack of same, of those making UDRP decisions.  Indeed the whole UDRP system ought to be reviewed by ICANN.  (Indeed, to my mind, the UDRP ought to abolished as the it has absolutely no relation to "technical coordination" of the net.  And as I see it, wee already  have enough lawmakers in the world; we don't need ICANN trying to be yet another legislature.)

Root Server Operator Memorandum of Understanding

We were presented with a model Memorandum of Understanding that is intended to be a skeleton for understandings with root server operators.

I voted to accept this resolution - I figured that having an "understanding" was an improvement over the current vacuum.  However I have significant reservations both about the vehicle of the MoU and its content.

I find the content of MoU to be an anachronism - it reflects the supposed halcyon days of the Internet's myth laden past when people did jobs just because they had a computer, a net connection, and felt like doing it.

The MoU represents an unwillingness of ICANN to come to grips with the fact that the Internet has evolved beyond those days.  The net has become a worldwide utility and its primary infrastructures need to be brought into predictable management regimes.  ICANN should be moving towards the day when it will run all of its root servers directly by itself or indirectly under clear operations contracts.  The MoU is an attempt to live in the past and to avoid taking the first steps along the path that will lead to maturity of DNS operations.

Let's recognize that the phrase "Memorandum of Understanding" is a phrase of multiple meanings.  In its best use, an MoU is simply a documentary device that records the current beliefs of two or more parties who hope to reach a legally binding agreement but who recognize that they have not yet done so.  In another use, "Memorandum of Understanding" is a weasel-word phrase used by those who are afraid to clearly acknowledge whether they are, or are not, willing to stand behind their promises.  Some MoU's are fully enforceable legal contracts, some are not.  This DNS root operator MoU leans towards the former, but is unwilling to firmly acknowledge itself for what it wants to be.

Certainty and predictability in DNS root server operations is a necessity.  DNS root servers have a direct and immediate link to the stability of the net as perceived by the users of that root.  If the servers composing a DNS root stop responding or start sending incorrect data, there will be a direct and immediate result.  Many people will be unable to obtain internet services.  Yet despite the direct link between root servers and net stability, ICANN is proposing the ambiguous and weak vehicle of an MoU.  This is in stark contrast to the situation with DNS registrars - registrars who could cease business today and the net would hum along without missing a beat - who ICANN has bound hand and foot in an iron web of clearly enforceable contractual chains.

This MoU insists that it's terms are binding, but the only remedy for violation is to terminate the MoU.  This is not very reassuring - it's sort of like a guarantee that the airline will refund the price of your ticket if the airplane crashes.

Don't you think that when it comes to something as important and critical as DNS root server operations that there should be absolute clarity?  Should there not be clearly drawn roles and lines of accountability?  Shouldn't the DNS root operators be at least as strongly obligated as the DNS registrars?  Do we really think that one of the most critical single points of failure of the Internet ought to continue to be operated as as an informal club of people who are unwilling or unable to stand behind their tasks and make a commitment to the Internet users of the world by saying "I am responsible, I am willing to be held accountable."?  The day for informal relationships at the core of the Internet are past.  Good intentions are no longer sufficient.

Which raises the question:  Why is ICANN choosing to use an ambiguous "memorandum of understanding" model for its all-critical DNS root operators?

If a DNS root operator is unwilling to enter into a clearly binding obligation then ICANN should thank that operator for its past services and seek either a willing replacement operator or should undertake to establish a root server itself.

With respect to the latter, I asked ICANN's management for an estimate of how much it costs ICANN/IANA to operate the root server that it operates.  ICANN's management was unable to come up with even an order-of-magnitude estimate.  I found that to be an appalling failure on the part of ICANN's management that strongly suggests that ICANN may have inadequate internal systems to identify and account for costs and expenses.

Have any of you seen this proposed MoU?  I do not believe that ICANN posted it for public review or comment.  Why the secrecy?

As for the agreement itself, there are terms in this agreement that I see as good and bad.

For instance, the MoU requires that a root server operator, should it cease to operate under the MoU refrain from having a DNS service at the old address.  This recognizes that the addresses listed in "hints" files are a kind of momentum:   Even if a root server ceases its role, resolvers run by ISPs, organizations, and users around the world will, for years to come, continue to send queries to the old address hoping to find the answer to the essential DNS question: At what addresses may I find a set root servers?  The answer that is given, if any answer is given at all, can capture the resolver and bind it to any of the various DNS root systems that exist.  That is a role that many would like to have as it represents a very distinct, very powerful, and possibly very lucrative point of control over the Internet.  And if no answer is given there will be timeout delays - perhaps imperceptible and incurred only at resolver startup.

The MoU requires the operator of an ex-root to convey to ICANN that block of address space in which that server was embedded.  Such a transfer is not only antithetical to the CIDR concept but its likely to engender hostility and even refusal from ex-root operators, particularly if ICANN didn't obtain the clear consent to this part of the MoU from the various organizations that control the address block and that may consider the block to be a rather valuable asset.

Perhaps the most significant aspect of the MoU is that it establishes ICANN as the source of the root zone file that each root operator would use.  This has several implications, one being new costs for ICANN to bear for the preparation and maintenance of that file - costs that ICANN's management were unable to estimate (again calling into question the status of, or even the existence of, ICANN's internal cost accounting systems.)  But cost isn't the most important aspect of this designation of ICANN as the provider of the root zone file.

With ICANN being the source of the root zone file used by the dominant system of root servers, one more prop of control by the US government is knocked away.  One wonders whether this provision of the MoU effectively renders vacuous the recent statements by Assistant Secretary of Commerce Nancy Victory that "the Department of Commerce has no plans to transfer policy control".


Return to Diary Index

Return to main page.


2002 Karl Auerbach, All Rights Reserved.
Updated: Wednesday, February 06, 2002 12:07:42 AM