Country code chiefs go nuclear on ICANN
Won't play with DNSO
When we suggested last November that
the country code registrars could withdraw from ICANN, we commended it as a threat to be taken seriously.
Such a secession would deal a terminal blow to ICANN's authority as the sole domain name policy maker, and one national registrar described it at the time as the "nuclear option." For morally, ICANN couldn't continue without the backing of the national registrars, and financially, these ccTLDs provide much of ICANN's working income.
Well, the finger moved closer to the red button at ICANN's Stockholm meeting this week, with the ccTLDs withdrawing from the key 'policy making' forum the DNSO, ICANN's Domain Name Supporting Organisation. (We hedge the term in quotation marks because it should be such according to the letter of the charter, but with ICANN being ICANN, it's more of a talking shop.) It's one of three foundation member groups of ICANN policy, the other two being the ASO, the Address Supporting Organization which looks after the numbers, and the Protocol Supporting Organization (PSO), which looks after the techie infrastructure issues represented by the IETF, the W3C and telco reps. So the country code chiefs aren't so much storming off carrying the ball, as much as they're disdaining to sit this one out on the sidelines in definitely.
This dramatic gesture could be construed as a step towards complete withdrawal, but while the ccTLDs remain inside the tent, pissing out, it can also been equally seen as a bid for increased status within ICANN. Interactive Week, has spoken to the heads of the ccTLDs including the UK's Nominet chief Bill Black, and the report coincidentally raises the issue of ccTLDs gaining seats on the ICANN board. Which seem too much of a coincidence for us, as the decision of the country chiefs to withdraw from the DNSO is such a powerful vote of no-confidence in the organisation. So, if you're really serious about leaving, then er, why do you want to sit its board?
ICANN's board composition gives three votes each to the three 'supplier' interests represented by the organisations listed above, but it also gives another nine member seats to the 'consumer' side of the Internet (the people who have to use the infrastructure, connect the networks, and buy the domain names)... with the casting vote in the case of a tie retained by ICANN's Maximum Leader du jour. That was the theory (and the promise). But of course four of those nine 'consumer' votes are retained by hand-picked ICANN placemen who haven't been elected by anyone. So the ccTLD's board representation could either come at the expense of the four stooges, or by lopping off even more seats from the at-large elected interests.
Now you'd think that there'd be hell to pay if ICANN did the latter, but perhaps it's actually not so difficult. All ICANN needs to do is redefine "consensus" in the language of kindergarten Marxist economics - which we have helpfully outlined above (as it's the only economics we know and trust) - to argue that consensus doesn't really need the input of more than a token of the hoi polloi. We can easily imagine the At-Large study concluding that the DNSO/ASO/PSO side of the table stays as it is, but the telcos and country code reps grab a portion the nine (rightfully) At-Large seats. Leaving perhaps a very token rump indeed of two elected representatives. Perhaps the only fig leaf ICANN feels it needs?
A pessimistic Michael Froomkin, writing at ICANNWatch seems to think ICANN can ignore the consensus issue: " My guess is that ... when it does get around to it, ICANN takes the position that the DNSO, and certainly not the GA or the non-member membership doesn't need to be consulted on this," he writes.
Readers who think the little scenario that we've sketched sounds implausible, should check the recent history of the British Labour Party. A history of the party's policy-making forum the NEC from the early eighties shows the same arguments being made, and carried, with an almost supernatural symmetry in the numbers. And we're sure you can find more examples... ®
Sponsored: Hyper-scale data management