Another interesting piece of information on states and private actors in the governance of the Internet. One of the more important parts of the Internet’s architecture is the management of domain names, like amazon.com, crookedtimber.org, and the various national top-level domain names (like .uk for Britain, and .ie for Ireland). People who study this sort of thing have a vague sense that domain name allocation is handled by private actors in most parts of the world. “Not so”:http://www.unicttaskforce.org/perl/documents.pl?do=download;id=448 according to a new paper by Michael Geist, who has preliminary results from 66 countries around the world.
bq. The most significant finding of this global survey is that, at least at the national level, governments are currently deeply involved in domain name administration. In fact, contrary to most expectations, virtually every government that responded to the survey either manages, retains direct control, or is contemplating formalizing its relationship with its national ccTLD. This is true even for governments, such as the United States, that generally adopt a free-market approach to Internet matters. Given the near ubiquitous role of government at the national level, it should therefore come as little surprise that governments have begun to seek a similarly influential role at the international level where policy decisions may have a direct impact on their national domains.
Geist’s evidence also suggests that the closer a domain name authority is to the private sector, the less likely it is to be interested in public interest goals, and the more likely its main concern will be flogging off as many domain names as possible. This isn’t exactly surprising, but it does offer food for thought, especially given “Verisign’s push”:http://www.circleid.com/article/538_0_1_0_C/ further to privatize control of the domain name allocation system.
{ 8 comments }
Ghost of a flea 03.27.04 at 7:25 pm
So, private sector registrars do not mirror the public interest (whatever that is). I am whether this study gave any consideration to those national registrars managed by authoritarian governments. But then that is not the sort of violation of the “public interest” many wish to consider.
Phill 03.27.04 at 8:50 pm
The study attempts to measure ‘the public interest’ in a very peculiar manner. Nine aspects of the public interest are proposed, one of which is ‘the public interest’.
When you try to tie down what those aspects might be you find that they are really two categories, things governments traditionally worry about and things that technologists generally worry about. It is not surprising that governments would be more interested in intellectual property issues than in technical reliability. But it would only take a single outage in dotCom and that perception would swing round instantly. Reliability would be government’s top priority – government answers to those who complained loudest and last.
At a recent Aspen conference Esther Dyson told me a story that illustrates the difference between the public interest and government. They were holding the UN Internet conference and there were big speeches being made by world leaders. The conference center had two doors, one for ‘government’, the other for ‘civil society’. So they are holding this conference on the Internet but the government keep the two groups apart.
Bob McGrew 03.27.04 at 9:54 pm
It seems to me that “flogging off as many domain names as possible” seems to be the primary purpose of a domain name authority. What do I want in my domain name authority?
1) a simple, transparent, easy process for getting my domain name
2) inexpensive domain names
3) few onerous restrictions on what my domain name can be used for (save possibly IP constraints) At the very least, there should be some domain where I can register and do whatever I want with my domain name. (At the top-level domains, .com is such a domain, though I believe .org and some others aren’t.)
It’s unclear where “the public interest” lies in this. And since government registrars (who were most likely to claim to be interested in the public interest) had the most onerous requirements (residency requirements) and the hardest processes to get domain names, I’d argue that the public interest here is being used as a smokescreen for inefficiency and political concerns.
Really, reading that paper from the standpoint of an end-user, it seemed as though commercial registrars seemed to have the right priorities and did the right things, whereas government registrars would be much less pleasant to work with.
So what’s the problem?
Bob McGrew 03.27.04 at 10:01 pm
Also, the study was unclear whether, with regard to the US, they meant US control over the top-level domains (.com, .net, etc), or the .us domain. Since I believe the .us domain was mostly intended for use by K-12 schools, libraries, and the like (though it has expanded from there), it would not be surprising if it was more controlled by the US gov’t than the top-level domains.
nnyhav 03.27.04 at 10:26 pm
food for thought
Onomastication?
eszter 03.28.04 at 12:09 am
Bob, I think some of the problems come when there are disputes over domain names. I actually don’t remember how tied in the registrar’s affiliation is to the dispute resolution process (can someone clarify?), but that may be an area where public interest may have a harder time prevailing when the registrar is private. (Again, maybe the two are unrelated, I’m going to need some clarification on that one.)
There were quite a few domain name disputes in the late 90s. These have become less common (and I only followed gTLD cases), but an interesting recent example is that of MikeRoweSoft.com.
Phill 03.28.04 at 2:30 am
I read through the article again, the guy does not seem to understand the difference between the registrars (retail), the registry (wholesale/supply) and the regulator (ICANN).
The issues he raises are the result of a peculiar set of circumstances. ICANN passes itself off as the global domain name registry authority. In fact it is no such thing, its authority is limited in practice to making policy for the non-country TLDs and to a limited degree for .US.
The country TLD operators have a very different view of the role of ICANN than ICANN itself. In particular they do not consider ICANN to have the right to make policy for their domains. It is not surprising therefore that they have a greater interest in the public policy aspects of domain name management than the registries. If push come to shove and ICANN did something seriously derranged (cut .cu (Cuba) off the net at the direction of the US congress because some senator wants to win votes in miami is a commonly used example), well in that case the country TLDs would revolt. That by the way is the reason that the Palestinian domain (.ps) is there to stay, that and the fact that Jon Postel was devious when he wrote the country code rules.
The registries are in fact nothing more than contractors who have been granted the right to operate certain TLDs for certain periods provided certain conditions are met. In some cases the registries have the right to renew their terms.
It is not surprising then that the registries would have concerns such as reliability, low cost, equality of access, etc uppermost in their minds since these are the issues that the contracts are intended to address.
The non-country registries have zero direct interest in intellectual property concerns because ICANN rules make it clear that this is a matter that is exclusively the concern of ICANN.
The non-country registries are also responsible for running many of the country code TLDs on behalf of the appointed policy bodies.
To the extent the paper makes any sense at all it is saying that different types of body will emphasize different issues. The part that is nonsense is when one subset of the issues are arbitrarily deemed to be ‘more important’. The main complaint that Internet users have about domain names are cost and reliability. The costs are already negligible in almost all the domains. The only real differentiator is reliability, and yes there are some country code TLDs run in the manner suggested that have not had satisfactory reliability.
The paper does not make the obvious conclusion that the best way to make sure that all nine issues get the attention they deserve it is best to have a division between the policy side and the commercial.
Better to have people worrying about the lights staying on all the time than have the policy people end up in reaction mode.
aelph 03.29.04 at 3:03 pm
whereas government registrars would be much less pleasant to work with.
As someone who works with commercial registrars regularly, I read this and laughed out loud. I’d rather wait in line with the DMV than deal with Verisign and their ilk any day. I’ve gotten faster and more helpful service from the *IRS* than I have from these companies.
Comments on this entry are closed.