Most accessed TLD'sTo asses the impact which adding over 1000 of new domain name extensions will have on current systems, ICANN has asked Interisle Consulting to analyse the log files of one of the internet root-servers.

The review of the logs revealed that a surprisingly large amount of requests is for domain names under extensions that actually do not yet exist, but that might be created in the future. Most requests where for .com en .net, but in third place there already is the first non-existing extension: .local. This even received over twice as much requests as .org. Problems with .local were however already foreseen as that extension has been blocked from being requested.

But already in 5’th place of all requests to the root-servers is the next non-existing extension; this time one that has been requested to be actually activated within the coming months/years: .home. Interisle reviewed a total of 96 hours of log files and during that period over 1 billion requests where received for .home. Other such examples are .corp (almost 150 million requests), .global (12 million requests) and .med (10 million requests). For all these TLD’s a request is pending to activate them, while clearly some systems currently expect them not to work.

It is not easy to predict the effect of activation such an extension that is expected not to work. And the volume seen also isn’t necessary related to the possible consequences. One of extensions that appeared in the test for example was .ice (almost 20 million requests), but when looking closer to the actual requests it turned out that most of them came from one company which clearly has misconfigured something causing such a large amount of invalid requests. Simply contacting them and getting them to fix this issue could solve the problem.

With .home, the scale of the impact could be larger, but maybe also not very serious. Such requests probably come from home networks and small consumer routers and activating .home as an actual extension might prevent your TV from being able to read movies from your network-drive until you install an upgrade. While this can affect many people the effect itself is not so bad as what could happen when it turns out that for example some heart rate monitors in a hospitals expect .med not to resolve to anything. Compared to .home only a small amount of requests are received by the rootservers, but it is easy to envision that the effect of one machine relying on .med to fail can have much larger consequences.

It is however very hard to predict what will actually happen. Just like with the Y2K-bug it might turn out that eventually everything does simply keeps on working. The figures only show which domain name extensions are receiving requests, but they don’t show why those requests are received.

It also must be noted that, due to how nameservers work and how information is cached, with the same amount of users requesting a valid domain name as users requesting an invalid name, more requests for the invalid names will end up with the rootservers. Because of this, numbers for invalid names are by design probably bloated over twice their actual proportion at the end-users.

There is no official news yet about the impact these figures might have on the timing and the release of any of these new extensions. ICANN however hasn’t ruled out that deadlines might change because of this.

Author

1 Comment