Lower your bow, ICANN

How digital archery threatens to become the organization's Battle of Hastings

Harold's demise as captured in the Bayeux Tapestry: is ICANN about to make the same mistake as the last Anglo-Saxon king of England?

Harold Godwin's army was already tired. The same day word arrived that Duke William II had landed on the English coast seeking the throne, the king's men had fought at the Battle of Stamford Bridge.

Determined to show he was in charge, Harold raced down to the village of Hastings, ignoring advice to let his men rest and to spend a day gathering reinforcements. Just a few hours later, he was dead, famously struck in the eye by an arrow; an event forever recorded in the Bayeux Tapestry.

The Battle of Hastings is one of the defining moments in the history of Western Europe, and one that could have been entirely different had Harold heeded advice. Likewise, a defining moment in the history of the Internet will soon arrive with thousands of new domain extensions. But having already fought a number of bruising battles, the organization in charge is insisting, against advice, on moving forward with a system called "digital archery" as a way to prioritize applications.

Under the system, all applicants will select a time some point in the future and then click their mouse as close as possible to the agreed time. The 500 fastest clickers (measured in milliseconds) will have their applications reviewed first. That process is expected to take between 5-7 months after which the next "batch" will be reviewed.

This approach has been complicated by a number of factors. For one, there are over 1,900 applicants, meaning the slowest clickers can expect to wait a full two years until their applications are considered - a lot of pressure for a single click of a mouse. Then there is the $185,000 per application that has already been paid to ICANN, which has lent an expectation that ICANN will deal with applications swiftly and efficiently. To top it off, ICANN is using the same system that in April revealed details of competing bids and forced a six-week delay in the whole process.

The situation is further compounded by the arrival of no less than four commercial services offering to increase your digital archery chances of getting into the first "batch" for between $18,000 and $25,000 a shot.

Predictably, the plans have been met with a wave of protest and formal letters. The process is "grossly unfair" and "flawed"; it will "create mayhem and ill-will"; and the organization is being "stubborn and inflexible". Those writing include long-term ICANN partners Neustar, NetNames and MelbourneIT, as well as the organization's intellectual property constituency. Two companies that will shortly become big players in the industry, ARI Services and Minds+Machines, have also written stark and critical responses. Weighing in is also brand adviser MarkMonitor, as well as a number of technical and domain resale experts.

Regardless, ICANN has insisted on moving forward and opened the digital archery system on 8 June - just days after it closed the application window - saying it will close the system 20 days later (the last day of its upcoming conference in Prague) and provide the list of names and batches two weeks later.

Why push ahead?

Of course, nobody purposefully creates a flawed system. ICANN sees digital archery as a way to tackle a number of separate issues: first, it had built all its plans around 500 applications and it wants to avoid disrupting the process by trying to review four times as many applications within the same process.

It has also made a commitment to Internet engineers concerned about the global network's stability not to add more than 1,000 new Internet extensions in any one year.

On top of that, it committed to governments that it would supply its advisory body (GAC) with no more than 500 applications to review at a time so that they were able to fit in with approval timelines.

As for the unusual system itself - clicking a button - that was devised by ICANN's lawyers, concerned that any system that did not include an element of skill could leave the organization open to running an illegal lottery (ignoring of course that most applications will have their "clicks" done for them by computers).

A juggernaut with blinkers

Critics - and there are many - argue that the current approach is based on assumptions that don't hold up.

The majority of new Internet extensions will be run on systems controlled by a group of around 15 highly experienced registry companies, negating the need for extensive and time-consuming technical checks on every application.

A large number of applications are also expected to be "dot-brands", or limited-use extensions, which would further reduce an application's complexity and hence the amount of time needed to review it.

As for the commitments ICANN has made, governments in the GAC just want to ensure they have time to properly review applications; the actual figure is secondary. This gives ICANN two options: extend the review timeframe, or work with the GAC to take uncontroversial applications off the list.

As for the technical limit, it is unlikely that ICANN will need to add more than 1,000 extensions per year due to the steps required before an application is finally approved, the likelihood of conflicts for the same name dragging things out, and the fact that companies will pick a launch date suited to their marketing efforts, rather than a mad rush to get out into the market.

But the problems go beyond questioning assumptions. Due to the way the process is currently designed, it is likely that the most contentious applications will enter the first batch - potentially bogging the process down in legal and procedural fighting from the start. It may also cause a rise in "defensive registrations" where companies feel obliged to register their trademarks under new top-level domains while they wait for their own name to be approved.

There is also a clear first-mover advantage to those in the first batch who will have a huge competitive advantage over those still waiting in the queue to be reviewed. It may cause those in later batches to drop out, sell on their rights or find different ways to make a profit. None of which is good for the program or the Internet.

And then there is the process itself: a game not of skill but chance. Even with advanced server technologies, it is matter of luck whether your response arrives within a window of 20 milliseconds due to the very nature of the Internet.

That inconsistent variation will be larger outside the United States or major Internet traffic hubs. Recognizing this, ICANN introduced a "round robin" approach where it will go to applicants in different regions in turn - but this last-minute fudge then puts North American applicants (where the vast majority of applications are expected to come from) at a significant disadvantage.

This high level of uncertainty has created its own mini-industry with four companies offering digital archery services to increase your chances of getting in the first batch. For a fee, you can jump the queue.

But even with the process off and running, it's still not clear whether ICANN considers using such a service legitimate or a breach of its terms and conditions (ICANN's refusal to provide an answer simply adding to the frustration).

And if all that wasn't bad enough, one applicant has published test results showing that ICANN's systems were introducing errors of up to one second - which could mean the different between being in the first and last batches.

As one respondent pointed out: "Another such 'glitch' in the earliest stages of the most ambitious and far-reaching project ICANN has ever undertaken would permanently damage the organization’s credibility, and likely call into question its continued viability as the steward of the domain name system."

Firing blanks

The same advice provided to Harold II on his way to Hastings has been given to ICANN many times: don't rush ahead, wait until you have all the information and resources you need.

For an industry that has seen literally years of delays, asking the organization to introduce more is a sure sign that something is very wrong. If there is time to review the actual applications (revealed later today), it may be that many of the issues digital archery was created to solve don't actually exist.

A number of alternatives to digital archery have been proposed: from batching by category, to identifying extensions that will bring competition, to prioritizing based on public interest. But the approach that has gathered most support is to scrap digital archery and batching altogether and run a single evaluation process.

Such an approach would extend the time it took for initial approval from 5-7 months to 12 months or so, but such a delay would be worth it to enable a level playing field and avoid the numerous problems that exist with digital archery.

One thing is for certain: ICANN's staff and Board are going to hear community views loud and clear at their meeting later this month in Prague.

If the Board continues on the path of keeping the organization's worst excesses under control, it will scrap digital archery altogether. If it decides to push on regardless, it will be one in the eye for all those that have worked for years to bring their plans to fruition, only to be repeatedly frustrated by the one company that has least to lose and most to gain.