Behind Closed Doors: Planning the Next Generation DNS?
Report on the First Meeting of the NAS Committee on Internet Searching and the Domain Name System
The first meeting of the new National Academy of Science (NAS) committee to study the Domain Name System (DNS) demonstrated the sharp contrast between the closed processes of the committee and the broad mandate from their US government sponsors. Those attending the one open session in two days of otherwise closed meetings were told that a reason this session was open to the public was because there is a legal requirement that a NAS committee cannot meet with its government sponsors in closed session. Though the study was requested by the U.S. Congress, the sponsors for the study are the National Science Foundation (NSF) and the U.S. Department of Commerce (DOC). At this meeting of the committee, the sponsors were invited to present the rationale and the need for the study and the issues to be emphasized or avoided by the committee.
The Internet is international in scope. It is used by people, institutions and governments around the world. It has been created through a process of technical collaboration. What then can a committee of people, many of whom have little knowledge or background in the processes that have made it possible for the Internet to develop and spread, do in terms of making recommendations for the future development of the Internet?
The presentations at this first meeting by the government sponsors outlined the rationale for the study. The US Department of Commerce representatives wanted the Internet to be changed so it can provide better protection for trademarks. Karen Rose, of the NTIA, raised the question of whether the DNS is a directory system or if "we're asking it to do too much." She told the story of how a few years before she had been told not to worry about problems with the DNS because there would shortly be a directory system, a better technology which could mitigate conflicts over trademarks. It was now 2001 she noted, and the DNS has not gone away. She asked the committee to cast their net widely to see what was available, abroad as well as in the US, to find a better technology that could be put in place.
Nobody posed the question of whether the Internet should be changed to better accommodate the needs of trademark holders. Nobody raised the problem of whether a general purpose communications infrastructure should be changed to accommodate the desires of a particular group of users. When a member asked if the committee should consider new addressing systems as well as naming systems, Rose said that these were related and so the committee should see what they could find out about what the implications are of the relation between naming and addressing.
Surprisingly, no one at the meeting explained the important distinction between naming and addressing. While the DNS maps to numerical addresses, it is not essential for Internet communication. The DNS could conceivably be changed or replaced without affecting the protocol that makes it possible to transport messages. A change in the IP addressing system, however, would have a serious effect on the technical operation of the Internet. The TCP/IP protocol suite requires unique addresses in order to route messages to their destination. The question of whether this committee should even be considering changes to the DNS is relevant when one realizes that there are technical organizations that are in charge of standard setting for the Internet and are responsible for the integrity of the technical development of the Internet. These include the Internet Engineering Task Force (IETF) and its leadership body drawn from the international community. Also, how the NAS committee's activities are to be viewed in the context of the functions of the IETF and other international technical bodies was not raised nor considered during the open part of this initial NAS meeting.
The NSF sponsors also made a presentation to the committee. One of the NSF representatives, George Strawn, presented a history of how problems administering the DNS had plagued the agency over the past 8 years. He explained that he wrote the solicitation that led to the agreement with Network Solutions (NSI) and which went into effect in 1993.
DNS registrations had cost the NSF one million dollars a year. They shot up to one million a month. A decision was made to allow NSI to charge for domain name registrations. The situation soon became highly political. The NSF held workshops to look at the long term problem in Winter '95, Spring '96 and Summer '96, but the workshops focused only on the short term issues. Strawn expressed the hope that this committee would look at the long term issues and promised copies of the workshop reports to those on the committee.
Along with the problems of scaling the DNS, Strawn urged the committee to consider the international nature and scope of the Internet. He questioned how it would be possible to balance a US centric and US government centric view with the fact that the Internet is a truly international infrastructure. He also asked what the role of government should be in the Internet's infrastructure. How would it be possible to morph IANA from a government supported entity to a true international structure? How close to government should it be? How far from government does it have to be? He described some of the lawsuits the NSF had been faced with. Since it was now 6 years since 1995 and the privatization of the NSF backbone, this was a good point to take stock and look at the trajectory. What would be highly volatile political problems in the future? He expressed an interest in the study supporting general discussion rather than being prescriptive. If the committee saw some real dangers ahead, they should say so. "Is there some train wreck coming down the road when the users of the Internet go into the billions?" he asked. He cautioned, "We don't want a melt down and so there is the need for practical solutions."
Becky Burr also made a presentation to the committee. She had been the US Department of Commerce official handling the development of ICANN. She noted that she, too, had been told not to worry about problems with the DNS since soon there would be a directory system that would resolve the problems. She described how some from the private sector had formed a group to administer the DNS but that they did not make it possible for other interested parties to participate so the US government stepped in. She did not mention the recent General Accounting Office (GAO) report issued about the relationship between the US Department of Commerce and ICANN. The report noted that there are constitutional requirements if the US government privatizes public property that is part of the Internet's infrastructure ("Under the Property Clause of the Constitution disposal of government property requires statutory authority. U.S. Constitution, Art IV, S 3, cl. 2" in General Accounting Office review of the relationship between the Department of Commerce (Department) and the Internet Corporation for Assigned Names and Numbers, p. 26). The ramifications of this report on the role of ICANN in administering the infrastructure of the Internet was not discussed.
Instead Burr's comments were concerned with how the committee would consider ways that technology can be made available to reduce friction for intellectual property owners and to protect consumers.
In response, committee members raised questions about how users would find information on the web. Never mentioned in the meeting were email and electronic communication like mailing lists and newsgroups and other forms of online collaboration between users to solve difficult problems, whether they be medical problems, or technical problems, or social problems, as important uses of the Internet.
The closed process of the committee was striking when compared with the open communication that has made it possible to build and scale the Internet. Not only was the earliest session on Monday, April 9 of the committee closed, but the committee's entire session the next day, Tuesday, April 10, also was closed to the public.
Even the open session was tightly controlled to keep the few public observers from having any way to contribute to the study. There had been little information about the committee available to the public. The few who came to the open session included an observer from the GAO and an observer from the French Embassy. Few from the press were there. Those observing were told they could not ask questions of the speakers. If they had a question they could write it on an index card and it would be given to the Chairman of the committee. Of the several questions given to him, however, the Chairman only read two. When one of the speakers had a question to ask the person who had written the index card, the Chairman would not allow her to answer.
Monday's committee meeting had another surprise. The final sessions were about what should be the criteria for judging alternatives and key principles. Some of the speakers were either present or former ICANN officials. The presentation by Mike Roberts, former CEO of ICANN, included a statement about the benefits of the Internets distributed, connectionless architecture whose utility, he said, had been demonstrated. He also explained that when ICANN makes a decision, its based on the sum of opinions of four or five major stakeholders. This helps to understand the contrast between the distributed nature of the Internets technology and the narrowness of the range and number of opinions determining ICANN decisions. Roberts presentation helped to point to the fact that some of those who are members of the NAS committee helped to create or supported the creation of ICANN. Yet the very creation of this NAS committee is an acknowledgment that ICANN did little to solve the domain name problem or the problem of Internet governance. In fact, the committee is a way to try to find a fix for problems that ICANN has only exaggerated.
One committee member was asked for his thoughts on what had happened at the first meeting. He responded that he would not be allowed to talk about what goes on in the committee outside of the committee. Also the list of future presentations to the committee was not available to the public.
Can a committee be expected to make constructive recommendations for the future of the Internet and for significant changes in the Internet's architecture and technology, when (1) its members have been chosen by some secret process (2) they have come from only one country (3) they meet in sessions that are often closed, and (4) they are prevented from communicating with the public? The Internet has continued to grow and thrive because of the open form of communication that is available online. It is the open processes that make communication and collaboration possible to solve the difficult problems of the Internet's development. The new NAS committee, however, is being carried out as a closed process, and yet it is being told it is to make proposals for big changes to the Internet, its architecture and its governance. Add to that that this committee is composed of people from only one country.
Clearly there is a mismatch.
It was good to hear the questions raised by US government officials that need answering from an international Internet research community. However, this committee is the opposite of such a community. One committee member asked a speaker: What global institutions are there that we might model that really work? There were no examples provided. But the nature of the question is revealing. The Internet itself, and the institutional forms that made it possible to conceive and develop it, are the models for the very form of global institution the committee member was asking about. The institutional form is that of bottom-up self-governance that has evolved as part of the Internet and that has made possible the collaborative development of the Internet. The institutional form making this possible was the Information Processing Techniques Office (IPTO) created at ARPA in 1962 by J.C.R. Licklider. The IPTO cooperative research process, first among researchers in the US, and then including researchers abroad, made it possible to give birth to the Internet and then to nourish its needed development. The IPTO was ended in 1986, but the lessons from its management processes continue to be needed for the Internet's development today.
When the Chairman of the committee introduced the open session, he gave a brief listing of some of the events that had contributed to the development of the Internet. During his summary, he included the ARPANET and the Internet as one continuous development, and the creation of the concept of gateways by Robert Kahn and Vinton Cerf as the development that made the connection of networks possible. At the end of his introduction the Chairman asked: Is the structure of the DNS robust enough to carry us into the future? Are there new structures? How can we find our way around the web? Are there new technologies? What is the institutional background?
The Internet, however, is not the web. Similarly, the main use of the Internet is not to find one's way around the web. The main use of the Internet is email and related forms of communication. Also the ARPANET and the Internet are not one continuous development marked by the introduction of gateways on the ARPANET. The Internet is not just a continuation of the ARPANET. The early ARPANET was a self-contained network that could not be connected to other networks. Other networks had to become part of the ARPANET, had to get permission to join. The early ARPANET protocol NCP and the IMP subnetwork were necessary to transport messages over the ARPANET.
The birth of the Internet was based on a new architecture, different from that of the ARPANET. The new architecture was called "open architecture" (A Brief History of the Internet). This architecture makes it possible for dissimilar networks created to meet different technical requirements and which are under different forms of political or administrative control, to communicate. The principles of open architecture have been embodied in the creation of the TCP/IP protocol. Networks adopting this protocol are able to communicate without having to become part of any one network or adopt the requirements of any one network. The principles of "open architecture" make it possible to federate networks into the metasystem we call the Internet. These principles are at the heart of what has made it possible for the Internet to spread around the world. A committee preparing a study to provide advice to the US Congress on proposals for future Internet development would need a firm grounding in what has made it possible for the Internet to develop, grow, spread around the world and scale thus far.
Attending and then thinking about the first meeting of the NAS committee leads to another observation. Some of the government representatives who spoke recalled how they were told there would soon be directory systems to replace the DNS. The creation of such systems is a research challenge. When such development is desired, it needs to be initiated and given support. The IPTO (1962-1986) provided the kind of institutional form to carry out such research. This is the kind of institutional form needed to create the technology to continue to scale the Internet. The researchers at IPTO worked collaboratively, first with researchers at universities in the US, and then with researchers in other countries to develop both the ARPANET and the Internet (The Birth of the Internet).
An understanding is needed of the institutional form that made it possible to create the architectural principles for the Internet and to design and create the TCP/IP protocol which embodies these principles. Such an institutional form can support the scaling of the Internet. The NAS study, to be a competent report, would need to explore such issues. The narrow composition of the committee and the closed set of procedures that they have adopted, however, are a serious barrier to being able to accomplish such a task. Licklider warned that when there is a difficult problem, it is important not to exclude the very sources and people who may make it possible to solve the problem. "There's a lot of reasons for adopting a broad delimitation rather than a narrow one," he wrote, "because if you're trying to find out where ideas come from, you don't want to isolate yourself from the areas that they came from."
Angebot des Monats:
Kaffee und Espresso aus Nicaragua in der Telepolis-Edition für unsere Leser
Warum man jeden neuen Emmerich-Film gesehen haben sollte
Mit dem Schalter am linken Rand des Suchfelds lässt sich zwischen der klassischen Suche mit der Heise-Suchmaschine und einer voreingestellten Suche bei Google wählen.
Zum Wechseln zwischen Heise- und Google-Suche
Verlassen und Zurücksetzen des Eingabe-Felds
Direkt zur Suche springen
Mit dem Schalter am linken Rand des Suchfelds lässt sich zwischen der klassischen Suche mit der Heise-Suchmaschine und einer voreingestellten Suche bei Google wählen.