March 13th, 2005

Innovation in DNS business

» ,

[ This entry is also on CircleID. ]

One thing that amazed me about the ICANN community is the creativeness in finding new business models. I am not even talking about new technology like Internationalized Domain Names (IDN), the number of business models created from the vanilla DNS (actually just .com) are just mind boggling.

ICANN was formed in 1999 and introduced the concept of registries and registrars model to the DNS business. With that, we witness the rise of, an IPO darling in the dotcom days, in the early 2000s and subsequently overtaken by the ultra-cheap high-volume reseller model of GoDaddy. We also see new registries like .info and .biz and several others that didn’t do so well.

There are also after-market (aka ebay) for domain names like afternic and registry outsourcing, DNS hosting, Dynamic DNS etc.

That’s about what most outsiders know of DNS business models, mostly revolved around the registry-registrar-reseller model. But there are really more and I shall discuss two not-so-well-known but interesting models below.
Read the rest of this entry »

March 3rd, 2005

.net domain name “tax”


Was scanning through some of my old comments and thought this conversation with Mark Mueller is interesting enough to be reposted.


Would be interesting to hear your view on ICANN’s intention to charge a per domain fee of $1 in the “.net” space? I believe they will not only start charging in the “.net” space, but also from new sTLDs. With all what is going on, it seems to me as the IAB (together with the IETF) will be next by charging a fee for each ENUM domain in “”. I don’t know who pays RIPE NCC’s expenses these days, but as we know we do not live in a “for-free” world… ;-)



I think it is inevitable that some part of the money collected will flow back to ICANN – be it a per domain name fee or something else.

For example, With a total expenses nearly 9M, ICANN must somehow have enough revenue. As indicated in the previous budget, it was proposed that the registries & registrars to pick up 8M of the tag but where does it comes ultimately comes from?
Read the rest of this entry »

February 28th, 2005

A Call for Resignation


This was posted to GA (General Assembly) mailing list by Danny Young:

I would like to begin by thanking Thomas Roessler and Esther Dyson for their service on the At-Large Advisory Committee and applaud their decision to resign in order to better pursue other endeavors.  I now call upon the remaining members of the At-Large Advisory Committee to tender their resignations. 

This is a vote of no confidence.  I believe that I speak for the At-Large community when I say that your service is no longer desired.  

The At-Large respects the concept of transparency; you do not.  No minutes of any of your meetings have ever been posted. No MP3 recordings of any of your teleconferences have ever been provided.  No record of any committee vote has ever been published.  At a time when the world expects the entities responsible for the management of the domain name system to be transparent, you have chosen to operate in secret.  It would be honorable for you to resign.

I have several discussion with At-Large committee members, at ICANN meetings, over emails, and also at the recent APRICOT.

I asked them what values they bring to ICANN beside signing up regional organizations which have little to do with ICANN charter, they provide no answer.

I told them that they are heading the wrong direction but none think so.

I offer to help them to use tools to facilitate online discussion, no response.

I have given up trying to change the current committee members. Hence, I fully endorse this letter (Thanks Danny) and call for all the existing At-Large committee members to resign.

January 27th, 2005

.net bids


ICANN is now seeking public comments regarding the .net bids. Unlike before, I am not going to offend one friend or another by siding with one proposal over another. They are all qualified and experienced registry operators. Instead, I will make some general observations.

1. None of the Revenue and Pricing Model (i.e. Section 4) about the bids are available to public. I wonder how the public is going to make judgement of one operator over another since cost is probably one of the few things general public would be interested to comment.

All the other stuff on DNS, Whois, registry system service-level etc etc can only be reasonable understood by an insider. Makes you wonder how “public” is the public consultation.

2. Highly politicise, with several top managements from companies urging ICANN to let Verisign keeps .net to prevent destabilization, I am surprised that there are bids that still comes in with less then 100% for DNS resolution of .net.

Anyone reasonable will tell you 100% is “impossible” but for DNS, it isn’t exactly that impossible given DNS is designed to be extremely reliable. You can have up to 13 servers scattered all over the network and more if you use anycast technology. And you tell me you couldn’t get at least one server running at any one time? (One is all you need to fullfill the uptime requirement)
Read the rest of this entry »

January 12th, 2005

Where did the .ORG money goes?


A friend pointed me to the latest Internet Society budget for 2005 :- ISOC is expecting PIR (ie, .ORG) to contribute 3.4M to the society! Wow, thats 2-3x as much as what Internet Society gets from its membership!

I think that’s pretty neat because ISOC has been in the red for many years and could certainly use some help financially. Afterall, it is hosting IETF and also paying for the IANA registry and RFC-Editors, all of which is critical to the Internet standardization process. Best of all, this also means ISOC would have some budget surplus to fund some useful projects :- 1.1M was allocated to that for Education and Policy development, including one doing African IDN :-)

But there is also another side to the story: Given there are approximately 3.3M .ORG names, this works out to be about US$1/name for ISOC. Hmm, didn’t each .ORG cost US$6 to the registrar so where did the rest of the money goes? Well, I suppose some money are needed to keep PIR going but PIR is a pretty thin organization1 so not much overhead there. Thus, I think the majority of the US$5 goes to Afilias, the registry operator for .ORG but discussion sake, lets just say US$3/name.

Now, back in the days when I was the CTO of i-DNS 3-4yrs ago, I remember doing a spreadsheet calculation on the cost to do full-fledge registry operation. At 100k domain names, it would work out to be about US4-5/name but once it reach 1M domain names, the cost falls to less then US$1/name. And it get sweeter as you gain economy of scale :- at 20M names, it cost US$0.10/name! So with 3+M .ORG names, I don’t think it would cost more then US$1/name for Afilias to run .ORG. That’s means a cool 7M profit to Afilias2 which is more then whats was contributed back to ISOC!

Now, I don’t have problem commercial companies making money. I mean, compared to Verisign which makes US$5.95/name for .COM, the profit Afilias makes is relevant little. But I think there is something fundamentally wrong when a commercial company makes money riding on a non-profit society. Shouldn’t more money be given to ISOC, or if ISOC don’t need it, why not lower the cost of .ORG registration and give it back to the Internet community?

This is also one of the reasons I don’t support the proposed .ASIA, with Afilias trying to replicate what the .ORG arrangement.I feel the AP community is been made use of and that feeling isn’t very good.

1 I only know Edward Viltz, the CEO & President of PIR but I haven’t met anyone else who works in PIR.

2 Disclaimer: I don’t have any insider information regarding the deal between Afilias & PIR so I don’t know how much they gets. Neither do I know the exact operational cost for Afilias so the 7M profit is just my intelligent (conservative) estimation.

December 19th, 2004

Engineers on TLDs


Engineers on TLDs: Do You Want Me With Fries? on CircleID by Vittorio Bertola:

This is why engineering solutions are not carved in stone, and young engineers have always been eating up old engineers since gave engineering to the world. That’s not because the old solutions were better than the new ones, or the opposite, but because new needs and new expectations arose among the users of the technology.

One word: Lovely!

I was one of the 30-something engineers who “ask” for (IDN) TLD.

While I can sense some consensus on stage and off-stage of the need of IDN TLDs, the response I get from 60-something engineers (whose status is close to been ‘God’ right now) revolves around trying to get things 100% right before proceeding.

Don’t get me wrong: I have great respects for elderly engineers who has contributed so much to make Internet what is today – and I always value their wisdoms and advises. But seeking perfect solution is no difference then say ‘no’ – at least to the one making the request.

December 10th, 2004

Adopt-a-TLD programme


Thomas Roessler responded to my previous entry on ICANN wrt to TLD.

Instead of being scared of possible failure, ICANN should think about how to deal with it. ICANN should be thinking about escrow solutions. ICANN should be thinking about how it can contribute to establishing a healthy and sufficiently large pool of operational registry businesses. And it should be thinking about how to help make sure that the registrations in any given TLD remain an asset in case of bankrupcy of a registry business, and not become a liability: Make sure that, when a registry fails, someone else has a business interest in picking up the remains.

Hear Hear! We need “adopt-a-TLD” programme for failed registries. :-)

December 6th, 2004



An anonymous writer posted an article title Time for Reformation of the Internet on Susan Crowford’s blog. The article calls for an liberal approach towards ICANN making lots of references to IETF and its process.

Two years ago, I might applause the article but since working for IDA, I learnt a lot of the policy making process to think otherwise. While I also called for ICANN to be more liberal and hence still agreed with some of issues discussed, I dont agree with the conclusion, particularly the references to Jon Postel. (See also Elliot Noss’s comment).

First observation: It is written by an engineer. As I commented to Joi & Roberto in Cape Town, it is a mistake to think engineers = good policy maker. At best, we are good advocates and advisors. I am not saying it is not important to understand the technology (and it is the duty of technologists to educate policy makers), I am saying policy formulations are far more complex, involving people, money and ecosystem.1
Read the rest of this entry »

December 4th, 2004

ICANN At-Large


I seldom talk about anything else other then IDN at ICANN meeting. But after several discussion over the week with various people, and prompting by Alan Levin, I decided to do so despite knowingly I am going to offend quite a few friends.

The topic I talked at the open mike during the public forum is on the progress with At-Large (ALAC). I was a long time critic of general public participations after witnessing how General Assembly attracted “colourful characters” like Jeff William. This remain so until Stuart Lynn, former president of ICANN who initiated the ICANN reformed, convienced me with one statement two years back: “Without ALAC, how are you going to contribute to ICANN as an individual?”

Fast forward two years to today, I believe something is fundamentally wrong with ALAC now. Not that I don’t believe at ALAC anymore but I am concerned with the approach ALAC has taken to gain community participation.
Read the rest of this entry »

December 3rd, 2004

Be more liberal please…


At the public forum, Elliot Noss (CEO of Tucows) brought up a controvisal topic of new gTLDs – Vint’s commented that DNS is designed to be hierarchical in nature and hence not able to support large number of TLDs.

To be exact, Vint is technically correct. And it is also true that we might have problems if we have large numbers of TLDs, creating a flat naming hierarchy and hence creating dependency on the (only) 13 root servers. However, such arguments are no long valid – we actually have more then 13 physical root servers, all scattered around the world via share-any-cast. Over 40 root-servers to be more exact and potentially as many as we want. So load-wise, we certainly don’t have problems supporting more TLDs then say 2 years ago.

What we don’t know is how many TLDs can we support now…nor do we know a way to accurately predict that. But my guess is 100,000 TLDs without a sweat.

The trend is that we already going flat. Restricting flatness in the TLDs means we get flatness in 2LD. (See related news on Verisign getting 5.1m new registration in last quarter.) We have over 66M zones under .com and .net and here we are arguing over a few TLDs?

While I don’t subscribe to Elliot free-for-all regime on TLDs, I think we can be more liberal with new TLDs.