Joi Ito's Web

Joi Ito's conversation with the living web.

In the last post I talked about wanting to break up monopolies and I was accused of being a hypocrite. The accusation was that ICANN was a monopoly. I responded by saying that ICANN is not a monopoly. One of the ideas behind ICANN was to break up the Network Solutions monopoly and encourage competition among registrars and registries. Also, if you want globally-consistent references, you need a root and an administrator of the root. ICANN is a non-profit and the board members do not benefit directly from the ability to regulate the delegation of top level domains (TLDs) and IP addresses. I think the trick is not to figure out a way to avoid anyone being administrator, but to figure out how to make that administrator fair. ICANN tries to address many of the issues by having a board composed of neutral members and members which represent the various constituencies. ICANN exercises a strict conflict of interest policy. Because of this, the board is very difficult to "capture" although a very broad group, such as the intellectual property lobby could feasibly have a great deal of influence over a number of the constituencies.

Anyway, having responded a bit defensively that ICANN is not a monopoly... and in the spirit of the soul searching from my last post, I do want think about what could make ICANN better. Even if I don't believe it is "a monopoly" in the sense of monopoly that I was talking about, it does have a monopoly over a particular aspect of Internet governance. I am going to Argentina next week to participate in the ICANN meeting so I suppose this is a good time to think about ICANN constructively and think about how I should try to contribute next week.

Just to frame this a bit more. I'm less concerned about what "evils" ICANN has done in the past and am more interested in what ICANN is doing now or should do in the future. ICANN is changing and fighting about the past is interesting, but not as productive.

A few things that I currently believe:

1 - The ITU can not manage names and numbers as well as ICANN and it's affiliated groups and I can't see anyone else who can.

2 - The basic architecture of ICANN - multi-constituent, multi-lateral with various working groups is correct.

3 - ICANN should and will eventually become independent of of the US Department of Commerce. The current goal is 2006.

4 - A completely distributed peer to peer directory service is technically feasible, but would be impossible to implement without causing complete chaos for people using the Internet today and isn't practical. Having said that, a more distributed directory system that sits on top of DNS may be useful, but that doesn't replace the DNS.

5 - ICANN should focus on names and numbers.

6 - ICANN should not become bigger than necessary to fulfill it's mandate.


"The ITU can not manage names and numbers as well as ICANN and its affiliated groups and I can't see anyone else who can."

Could you expand on this at all? I'm assuming you're refering to the IANA function here - a job which, previous to ICANN, was performed by one man - John Postel.

Given that ICANN don't manage nameservers, or manage IP allocation (beyond, I'd assume, allocating blocks to regional registries), and only registering the results of the IETF processes... does ICANN really represent the best, and most cost effective, way of doing this?

It doesn't seem to be the IETF, or ARIN/RIPE/APNIC (or anyone else that's actually doing the bulk of the technical and administrative side of running the net) that a lot of people have a problem with - it's *just* ICANN.

Not that I think the ITU would work out cheaper (and I wouldn't propose them as a candidate), but I find it difficult to believe that there's no way of performing this task without spending the mad money ICANN seems to require.

Joi:ICANN has tried to break the monoply of the Dept of commerce and IANA. Why is it taking so long ??

I have to agree with Lee, Jon Postel was doing that as well as being the editor of RFC's!! but this was before the internet become the World live web.

As such "ICANN established market competition for generic domain name (gTLD) registrations resulting in a lowering of domain name costs by 80% and saving consumers and businesses over US$1 billion annually in domain registration fees." Will its 1$ worth of market cap - and no wonder there is so much bickering -who wants to loose 1$B eh ??

Secondly, IPv6 has been in the pipeline for over 15 years. Yet, the community is not seeing a cohseive roadmap on how this resource will be assilimated, shared and deployed. the 6bone network is address space is being allocated thru IANA. Where is ICANN's roadmap for this important resource which the internet requires ?? Yes, its a mission that ICann is aspiring for, but yet to seen. How this will be accomplised ??

My understanding is the Jon Postel didn't do it all himself. He always had a few people around to help out. Having said that, ICANN is much larger than Postel in his team. I think the issues have gotten much more complex with governments, WIPO and their trademark issues, businesses running TLD screaming for more TLDs, engineers worrying about stability issues screaming for no more TLDs and a great number of people pushing and pulling on every issue. Having said that, I would agree that it is the board's responsibility to keep things as small and simple as possible. This may sound like bureaucratic process mania, but an important step is to go line by line through the strategic plan and really think about whether each thing in it is something that ICANN should be doing and if so, whether the proposed method or process is the most cost effective and efficient.

I guess this may sound like a weak argument, but I think that moving ICANN functions to the ITU would probably end up in a less efficient and even more difficult structure. There are some good arguments online about this and I can provide links later if anyone is interested. (I'm on gprs in Hong Kong airport so I'm going to refrain from a Google-a-thon.) I can't see how reinventing it from scratch would produce any better results. We're climbing a learning curve, even if it doesn't look this way on the outside and I still stick by the opinion that "fixing ICANN" while trying to keep it's growth in check is still better than any of the other options.

Good question on ipv6. I'm not an expert, but I'll try to learn more in Argentina.

Regardless I agree or disagree with your position (and btw, I agree with almost all of them :-), I applaud you for coming forward to state your position. This is something we seldom see in ICANN board.

Thanks James. So... maybe I'll kick myself for asking, but what do you disagree with?

Fact: IPV6 has very little traction.

A lot of people assume that it should have traction, but I'm not so sure myself.

The technical solution is well-bounded, accepted, etc. I don't think anyone would quibble with me stating that its ready for primetime. *But*, if no one is asking for it, what problem is being solved?

Not implementing IPV6 isn't really ICANN's failure. Certainly, we could take a stronger leadership position in the discussion, but sometimes, the best management is no management at all.

If ICANN is going to succeed in its mission, it needs to have a better understanding of this dynamic. It needs to understand when to step back and when to take an active role - when to implement policy, and when not to implement policy, when to encourage protocol adoption and when to stand aside.

"The ITU can not manage names and numbers as well as ICANN and it's affiliated groups and I can't see anyone else who can."

ITU afterall manages lots of numbers, from E.164 phone numbers to GSM code that is been assigned to every GSM phone. To claim they are unable to manage names and numbers is unfair to them.

OTOH, ITU have a way of doing things that "Internet people" are not comfortable - the burearacy, the government centric view to all. Fundamentally it is a culture crash.

Yet OTOH, we cannot deny two things: (1) Despite the uncomfort, ITU does get things done - things are slower but really, ICANN/IETF arent much faster these days either (2) "Internet people" may be uncomfortable but most other general internet users don't really care. And we have a lot more "general internet users".

IMO, it is always good to have ITU as an option, as a backup just in case. Or if anything else, a little competition to keep ICANN on its toes.

James, maybe you can give me some more information on this, but from my understanding and what I hear about what ITU people have said, they seem to think that experience with phone numbers somehow translates to experience with names and numbers. I think it has even been suggested that IP addresses get handed out to countries like phone numbers. I think you know better than me that IP addresses don't work like phone numbers. They should map network architecture, not national boundaries. It would be very inefficient and possibly impossible to allocate IP addresses in the same way phone numbers are managed. I'm not sure how well the experience with "numbers" that ITU has will translate into expertise with DNS or with IP addresses. In fact, I would assume it would hurt because the familiar methods are not applicable.

I'm not against ITU opinion. I'm just worried that they are "looking for the next big thing" and are trying to expand their mission.

Joi, I find it very hard to resist the "ICANN as monopoly" tag when ICANN has been preventing competition since its inception.

I don't need to remind you of the history of my application for .Web in 2000, how it was almost handed to another company even though we'd been involved in the process since day-one (working with Jon Postel and then the formation of ICANN itself), and how we've still been waiting to compete for almost TEN YEARS now.

Yes, it's been TEN YEARS since I first proposed .Web and built the registry as a proof of concept at Jon Postel's urging.

Yet ICANN continues to prevent me, and companies like mine from competing, while allowing registries like .PRO to be sold and co-opted from their original mandate. .NAME was supposed to be 3rd-level only, yet is allowed to sell generic 2nd-level domains. Sponsored TLDs are allowed to cut in line in front of the 40 or so applicants from 2000 who are still waiting, after having paid their $50,000 only to see new TLDs be given to industry insiders.

And now .eu gets to jump the line, too.

If ICANN isn't a monopoly in this control of the market, what is it?

1 TrackBacks

Listed below are links to blogs that reference this entry: ICANN?.

TrackBack URL for this entry:

The Register has a good coverage on why it has taken so long. I really don't know whether this is a good idea or not. I understand the Internet has developped from its early days as a military project in Read More