ICANN ICANN Email List Archives


<<< Chronological Index >>>    <<< Thread Index >>>

On "The Public Interest"

  • To: affrev-draft-processes@xxxxxxxxx
  • Subject: On "The Public Interest"
  • From: Eric Brunner-Williams <ebw@xxxxxxxxxxxxxxxxxxxx>
  • Date: Sun, 31 Jan 2010 12:07:22 -0500

This is offered as a response to Steve DelBianco's thoughtful effort to start a community discussion, the first note on this public comment thread at ICANN.

> The Affirmation of Commitments (AC) adds a new imperative: ICANN decisions must now meet a 'public interest' test.

RFC 1591, Domain Name System Structure and Delegation, edited by Jon Postel in 1994, contains this guidance on the administration of delegated domains:

"[designated managers of TLDs] are trustees for the delegated domain, and have a duty to serve the community.

The designated manager is the trustee of the top-level domain for both the nation, in the case of a country code, and the global Internet community.

Concerns about "rights" and "ownership" of domains are inappropriate. It is appropriate to be concerned about "responsibilities" and "service."

Of course, this guidance did not spring out of the earth, like crocuses, and apply only to delegations from the then-unique DNS root, but was organic to the IANA function, and to the successors-in-inteterest to Jon as the "Numbers Czar". I offer that had Ira Magaziner accepted the advice of those involved and adopted the IAHC model and the CORE of 1998, that "public interest" would be much more evident in the actions of the coordinating institution, but Ira's choice of a "private-public multi-stakeholder" ment that "public interest" has been present, if not predominant, in the institution of his choice in 1999.

On to issues.

> First, I believe that the 'public' part of public interest is concerned more with users and registrants than with contracted parties and others who are deeply involved at ICANN.

Agree, but with the caveat that users are, in the first instance, applications that attempt to resolve identifiers and thereby obtain routing to resources. We must not loose sight that the function exercised by the institution, Ira's choice or Jon's, is technical coordination. Also, at least two types of registrants, free speach" and "trademarks", are represented, and over-represented [1], respectively, in the Constituency (now Stakeholder) model of the institutional structure.

> Second, I believe that the public interest in ICANN decisions is somewhat broader than just a secure and stable DNS.

Agree. Presumably a mapping function at the level of transport, which provides routing information, has been created for some larger purpose than mere existence, which could be "secure and stable", or not. However, your next sentence is where we begin to disagree.

> Namely, users and registrants want ICANN to make sure the DNS delivers two essential and measurable qualities: *Availability* and *Integrity*

On "*Availability*":

We actually don't know what "users and registrants (in the abstract) ... want". We do however know what specific users and registrants want, and if we discard the perennial "trademark" and "free speach" statements of positions, what we are left with is diversity, of language as you note, and diversity of an ensemble of conditions that I will call "material diversity". By this I mean the appreciation that ICANN actually meets outside of Marina del Rey, and the desire that ICANN not merely meet outside of Marina del Rey, but that ICANN's values, and in an imperfect world, necessarily its staff, and also those with access to its staff, reflect the material diversity of a world larger than Marina del Rey, and larger than the early adopter world at the point of deregulation and commercialization in the chronology of the ARPANet and its successors.

If there is one thing we can be fairly sure that "users and registrants (in the abstract) ... don't want" it is that fabulous creature, "100% uptime, 24 hours a day, 365 days a year, from anywhere on the globe", as users and registrants are used to best effort as the fundamental technical service description, and do not choose soft, or hard real-time services for identifier-to-resource mapping.

As a rhetorical device "uptime" is a poor choice of foils with which to attempt to skewer ICANN. As I write this Bank of America has pretty much dropped off the net, which makes me smile for a variety of reasons, though I am a depositor to, and my primary financial service provider is, Bank of America. Users and registrants (in the abstract) ... "route around the damage" or wait, whether the damage is government censorship or severed cables.

On "*Integrity*":

An example we at CORE have used over the past few years to motivate what we think is the better sense of consistency of meaning is the domain "whitehouse.us". Up until recently, it resolved to the ppc site of choice of the NXDOMAIN monitizing ISP providing network access.

The fundamental assumption or shared truth of the DNS is that people can use textual identifiers -- strings -- with more utility than the dotted quads that are the string representations of the 32bit IPv4 addresses. An attractive side-effect is that operators may renumber as their address requirements change without users being aware of, and forced to change their use of the associated resources, through the use of "names" as identifiers. Further, these identifiers are not private, they are not in the secret language of the mind of each individual using the DNS, they are public and broadly shared. Users have expectations about identifiers they do not create, yet use.

Oddly enough, it is the pay-per-click business model which is most illuminative of this shared fate of identifiers. PPC does not have a random distribution in the [a-z][0-9][-] string space, nor does it simply form small "Hamming Clouds" around trademarks, it also "squats" on commonly used though not "generic" identifiers. PPC "shows us" what our common expectations about identifiers are.

An earlier model was the "city-squatting" of the .us name space during the IANA tenure, unfortunately retained, though renamed, during the NeuStar tenure. However, not only do we expect New York to exist, and Las Vegas to exist, we expect that within each of the name spaces the common American taxonomy of libraries, museums, cooperatives, hospitals, clinics, airports and transports, ... all exist and exist consistent with the expectations of frequently resource resolving users.

In CORE we call this the "Name Space Mandate". The details of script and taxa are particular to culture, obviously what works for Paris is not identical for New York, yet the structure, and the supported set of expectations within an identifier cloud, are common to both.

This, and not "accurate and authentic results" when they attempt to purchase some widgets at widgets2010.example, is a more inclusive sense of "integrity". The use of the net for informational purposes is orders of magnitude greater than the use for commercial transactional purposes. We know this because of the vast amounts of money spent attempting to convert "informational eyeballs" into "transactional eyeballs" and "clicks" converted into "sales".

At present we have a theory of marks which makes some expectations as to the resource associated with trademarks rational. We also have a theory of policital speech which makes similar expectations rational. We have nothing else except the admission control policies of some registries, and some registrars, to support expectations about the resources associated with names in .coop, and .museum, and .cat, and so on.

However, even the expectations of users as resolution initiators for either "availability" or "integrity", alone or together, is insufficient to meet the best sense of "public interest" in the present moment. The authors and operators of the srizbi, rustock, mega-d, hacktool, pushdo and storm botnets, and their commercial customers, act contrary to the public interest.

The manditory implementation of BCP38 is in the public interest. The abandonment of 15 second TTL as the default by flux supporting ("rapid update") registries is in the public interest. There is a non-trivial shopping list of things that are profoundly overdue and which are in the public interest.

In closing this already lengthy note Steve I am glad you encourage further community discussion on the concept of public interest in Affirmation reviews. I agree this term is too important to leave undefined or let a few reviewers define it to fit their own agenda.

In "The Public Interest" must mean something, we already have trouble with the meaning of "security and stability" being rather elusive.

In a personal capacity, though I appear to have slipped in an advert for CORE's work on urban places and names.

Eric Brunner-Williams

[1] A correlation presented at the Paris meeting showed that neither of the ISPC nor the BC could be distinguished meaningfully from the IPC on the basis of their votes on GNSO Council (policy) questions.

<<< Chronological Index >>>    <<< Thread Index >>>

Privacy Policy | Terms of Service | Cookies Policy