The common misconception is the Internet is all de-centralization. It is actually both; Above IP, it is and should be absolutely decentralize but when you dealing with core infrastructure (IP & DNS), like it or not, that’s is centralized.

Another misconception is ITU is all about the ‘old worlds’ and didnt ‘get it’. Regardless of the history, look at some of the issues that has been discussed in the recent WTSA and it is obvious they did get it.

Back to topic, there are some fundamental problems with ICANN. Here are some of what I think:

1. Building trust among stakeholders – particularly governments around the world is extremely important yet lacking. The reality is that many governments play a big role in developing Internet in their countries (especially outside US) and not engaging them is silly. Transparency is important to build trust, not the other way round. ICANN been a US corporation is also not helpful to towards this end.

2. Speaking of trust, it is also related to the arcane root server operation. Be more open about it – that all the root servers (including those shared-anycast ones) takes the zone file from “A” Root Server, which is operated by Verisign. Verisign in turns takes the zone file changes order from ICANN/IANA signed off by US DoC. The US DoC link is bad. Hiding it wont make it go away.

3. ICANN is bloated; for the amount of work they are chartered to do, they have too many highly paid executives doing too many travelings and too many committess bloated by lawyers & etc. They should look at ways to cut-cost and ‘delegates’ more to the community. All they just need to do is ask and RIRs and regional Internet organizations would be more then happy to help. Not engaging these existing organizations and having more committees of their own only increase cost yet adding little value to the whole.

4. ICANN have not found a good mechanism to engage the at-large community. Excerises such as the (GA) @large mailing list, the @large committee etc produce very bad signal-to-noise ratio. Hopefully they could start looking at blogging. Already, efforts like CircleID show the potential. CircleID produced much better and more constructive inputs from community then I have seen for years.

5. Releasing more TLDs, especially IDN TLDs (i.e. non-English TLDs). The failure of some TLDs does not indicates we need less. Innovation is a number game…1 in 100 at best. You want to be successful, you need to be more open minded. Don’t be a innovation-bottleneck, or believe you know which TLD will be successful and which will not.

6. Streamlining operation on DNS transfer, especially ccTLD transfer. And on this topic, there will be instances where the ccTLD operators are at odds with their own government. Like it or not, ccTLD is considered a ‘national asset’ for many government (see 1) and ICANN needs to accept that. Failure to do so will only push governments to the venues who will listen to them, venues like ITU.

7. Last but not the least, WSIS. I have just one advise: Focus on doing your job well and no one can take it away (or at least, you wont spend too much time to fighting to keep it). The only time you need to engage in serious politicial games is when you are not doing your job well.