Re: [ga] Re: alternative root DNS systems which resolve ICANN domains as well?
Stephane Bortzmeyer wrote: It cannot work because you need a way to resolve collisions like ".biz" or ".home" (which appear in several namespaces).
As long as each root system provides the same set of delegations (NS records) for each TLD then there is no problem for those TLDs that the roots have in common. We have a reasonable degree of operational experience in this regard via ORSN and the less orthodox competing roots that do honor and copy the NTIA TLD NS record delegations. The concerns, and I raise them only to the level of concerns and not problems, appear in two regards: - A) When the sets of TLDs offered by different TLDs do not exactly overlap. For example, some root system may offer boutique TLDs that are striving to obtain market share while the NTIA (we now know that it's neither ICANN's nor IANA's) would contain only the 250 or so TLDs that most of us know about. But, and here's the important part: among all the root systems they honor the same delegations (NS) to the same named boutique TLDs. - B) When sets of TLDs offered by different TLDs do not exactly overlap and the delegations to same named TLDs are different. Generally case A is considered benign while B is considered very bad. Why A and not B? Case A is a general mechanism for market-driven expansion of DNS without all the micromanagement of ICANN. It let's the market decide which boutique TLDs will rise to prominence and thus eventually appear in most the catalog of offerings of most root systems. Some people are concerned that this will break "universal" connectivity on the net - I agree that it may mean some people might have to go to a different root zone to find some boutique offerings, but that's the price of a market-driven selection system. And besides, it's part of our normal everyday life in other areas: We don't go to the neighborhood supermarket to find really nice brands of Champagne, we go to a specialty store. But market demand has pushed all the larger brands onto the shelves of virtually every supermarket. Moreover, the idea that DNS names are permanent is something that is simply not true, they change over time. And URL's/URI's (which often contain domain names) often vary with location and querier as well. And email addresses rot nearly as fast as cantaloupes in summer. Given the erosion of DNS names, URLs, and email addresses over relatively short periods of time, I'm not sure how much weight to give it as a reason to nail DNS to ICANN's immutable iron cross. In case B, fights over who gets to have what boutique TLD name can be fought out in the well-established and well respected fora used for trademark and service mark disputes. The idea behind case A is that if the TLD exists and can be reached, it has the same contents. Case B is generally considered bad because it creates situations in which users, software, administrators, and troubleshooters can be confused. There is no technology on the net to prevent either Case A or Case B. And to impose such a limitation would, at least in my eyes, be at odds with the end-to-end principle that places user choice at the edges of the net. Whether or not we like it, ICANN's resistance to new TLDs has created a large pressure to break out of the current ICANN stagnation. One of those ways in which pressure may be released is through the creation of new root systems or, as seemingly happened in China, some large areas have added additional nominally locally-visible TLDs to the servers used by their users. --karl--
|