Alternatives to U.S. Internet Governance

October 11, 2006 by admin

Wolfgang Kleinwächter

Wolfgang Kleinwächter

What alternatives are there to extending the ICANN agreement with the U.S. Department of Commerce?

Kleinwächter: As long as no new, functional system exists, we’ll have to live with the second-best solution: control by the United States. In June 2005, Kofi Annan challenged the UN Working Group on Internet Governance (WGIG) to be just as creative with the discussion of political questions of Internet governance as the developers of the technical infrastructure were in the 1960s and 1970s. We’re still waiting for an innovation. Transferring the Internet to an intergovernmental agency in the UN would be the worst of all possible options. If ICANN presented itself as a transparent organization that acted in the interests of and in cooperation with the global Internet community, but remained independent of the government, we would be one step ahead.

When could U.S. control of the Internet end?

Kleinwächter: There are two contractual arrangements of note here. First, the general Memorandum of Understanding between ICANN and the U.S. Department of Commerce has a time limit. It runs out when ICANN is “mature” enough for independence. That means when it has implemented a transparent and reliable decision-making process and has developed a sustainable foundation for relationships with the regional Internet registries responsible for IP addresses and with the root server operators. As an intermediate step toward independence, ICANN and the U.S. government replaced the memorandum with a Joint Project Agreement at the end of September 2006. The Joint Project Agreement lengthens the leash between ICANN and the Department of Commerce. The agreement is valid until 2009.

The IANA Contract between ICANN and the Department of Commerce regulates management of the root zone files for top-level domains (TLD) and is more complex. The TLD root zone files are managed by the Internet Assigned Numbers Authority (IANA), a unit of ICANN. IANA controls and modifies the files; it can also delete or insert new TLD zone files. But before the TLD zone files are published, IANA must request approval of the data packets from the Department of Commerce. That’s the trigger for the initiative.
There are plenty of suggestions for alternate methods of authorizing TLD root zone files. But most of them involve replacing one unacceptable, but functioning system with another unacceptable – and as yet unproven – system. Authorization usually involves purely technical considerations. In my opinion, ICANN should develop procedures that involve authorization by a governmental agency only in cases with political aspects. In such cases, governments can create ad hoc working groups or involve ICANN’s Governmental Advisory Committee (GAC). The eIANA procedure for automatically modifying ccTLD zone files that was developed by the Polish registry NARSK is a step in the right direction. No individual check occurs; the file is practically waved through.

Have there been any cases of manipulation or misuse by the American agency?

Kleinwächter: Some people are quick to assume that the U.S. government could simply disconnect a country – a rouge state or a disloyal ally – from the Internet. No such case has become known so far. Even media savvy blackouts of domains like .af (Afghanistan), .ly (Libyia), or .iq (Iraq) turned out to be issues of commercial conflict. In most cases, rival groups in each country could not agree on who should manage the country’s domain names. The U.S. government actually could have ordered the VeriSign security firm to remove a country’s TLD zone files. But aside from political damage to the United States, the effect would be rather minor. Since the creation of the Anycast system, there are more than 100 root servers. If only one of the servers does not execute the delete command for the ccTLD zone file, communication with the ccTLD can still function – even if there’s a bit of a time delay because of longer communication paths.

What speaks against the UN taking control of Internet governance?

Kleinwächter: Can you imagine that the United States, the European Union, China, India, South Africa, and Brazil – six potential candidates for a UN Internet security council – could ever agree on uniform criteria for freedom of expression? No. Supervision of the Internet should remain with the technical developers, users, and Internet service providers from private and civil society. They’re all in the ICANN boat.

What’s the goal of the newly founded Internet Government Forum (IGF) at the UN?

Kleinwächter: The IGF was conceived of only as a platform for discussion. It should not make any decisions, but simply support communication between the parties involved and develop new models. In principle, the IGF itself is a new model for developing policy. Decisions should be made in a decentralized manner. That tends to make the mechanism of political decision making similar to the technical infrastructure of the Internet itself: a network of networks without a center. Whenever a question of power on the Internet has arisen, it has been given to end users. The IGF can drive this process further ahead. GigaNet also wants to contribute to this goal.

Is it conceivable that several parallel Internet worlds will exist at some point?

Kleinwächter: All TLD root zone files – those with .com – have so far been based on ASCII code. And the file for a TLF exists only once. The Chinese variant of .com is a new root zone file. And that’s the question. Who authorizes publication of this new file? VeriSign claims a type of universal trade mark for .com and .net in all language variants. PIR makes the same argument for .org and Afilias for .info.

But the Chinese have clearly stated that publication of TLD root zone files with Chinese characters cannot depend on authorization from the U.S. government. During the current tests in Beijing, the Chinese name servers add the ASCII ID of .cn (the country code for China) to the Chinese versions of .com and .net. That’s how you stay in the ASCII, IANA, and ICANN system. But the problem is that any Chinese in San Francisco searching for a .com address with Chinese characters won’t find it unless they know that they must add .cn to the Web or e-mail address. For the Chinese, it’s a simple matter to stop using the ASCII supplement of .cn and set up their own root with their own root servers and root zone files with Chinese characters. The publication of these root zone files could then be authorized by the Chinese Ministry for the Information Industry.
Alternate roots with their own TLDs are actually nothing new, but they have so far failed to find a critical mass of Internet users. But that will change with almost 500 million Internet users in China. So if there is to be a new root, the question arises of how both roots are linked with each other. Everything can be resolved, but no one quite knows how to do so yet.

Can an initiative like the alternative Internet in Turkey, which is not controlled by the United States, threaten the idea of a single, global Internet?

Kleinwächter: I regard such an alternate – or actually a complimentary – root as realistic only for languages whose alphabets are not based on ASCII code: Russian, Greek, Person, Arabic, Korean, Japanese, or Hindi. In fact, these languages are the majority of languages listed in the ISO 639 standard. For languages with Roman letters, such as Spanish, German, French, or Czech, that simply contain additional characters not covered by ASCII code, having a distinct root makes no sense.

It also makes sense to set up an individual, language-based Internet root only where there is a critical mass of users to avoid stewing in your own juices. The Chinese can live with two networks. The Chinese network would then be used internally. And the 5 percent of the population that wants to or must look at the outside world (which is still 80 million users) could receive a government-issued password that enables access to the ASCII Internet, which is administered by ICANN and IANA. The same approach could work with Arabic, although there is no uniform language table for Arabic. Russians and Iranians might well have a political interest to protect themselves from the outside world. It might also work for the Koreans and Japanese, but I don’t see any political interest there in going it alone.


Leave a Reply