Introduction

The "privatization" was undertaken without any agreement among the players of what would constitute criteria for success, either in terms of network access and cost; in terms of technology and connectivity to other networks; or in terms of usability -- would people be able to do with the network what they want and need? Indeed one wonders whether what has happened should even be called privatization. Consider the following comment: "Privatization and free markets don't result from exclusive government contracts for $10 million a year. The free market results when the government gets out of the business, and lets individual buyers purchase whatever networking they are willing to pay for, from whoever they want to buy it from. Privatization means that the government no longer provides a service, not that they subsidize a private company to provide the service."¹ and ^a

On November 1, 1992 policy at the NSF had come to such an impasse that the agency - by administrative fiat - extended Federal payments for backbone provision to ANS for another 18 months.^b This in spite of the fact that both ANS and NSF documents released to the author by the NSF under FOIA a month ago indicate that the NSF has had some difficulty with both MERIT and ANS in enforcing the conditions imposed by NSF when it agreed to the "privatization."^c Apparently the NSF is marching onward because the only other choice would be to terminate the cooperative agreement with MERIT and watch the structure of the NSFnet collapse.

The NSF is paying ANS $10 million a year for a backbone that is supposed to operate at 45 megabits per second (mbps) but in reality operates at closer to 8 to 10 mbps. The NSF is also subsidizing, at a rate of approximately $7.5 million a year, a mid- level network infrastructure that connects 1200 colleges and universities to the backbone.² and ^d ANS is functioning as a direct competitor of this mid-level infrastructure by acting (in telephone industry terms) both as an inter-exchange carrier and a local exchange carrier. ANS has told the NSF what it intends to do to develop and expand the network. But because ANS now functions as a private company with a for profit subsidiary, it is by no means certain that the NSF will be informed whether ANS game plan is succeeding in a timely enough way to allow for informed policy making by the NSF. Meanwhile there are signs that ANS will have considerable difficulty in meeting the goals it has set.

While the US government still uses taxpayer funds to build up the mid-level infrastructure, and still uses taxpayer funds to buy access from its commercialized backbone provider, it has apparently given up all control over the network backbone.^e In doing so, it has given up its ability to define the network in terms of the public interest. Furthermore while the NSF remains heavily involved in funding the mid-level networks, it has also acted in such a way that the outcome of events affecting the backbone could force it to significantly increase mid-level subsidies. As this report will show, ANS, into whose hands the NSF has placed the backbone, appears to be a competitor of the mid-levels having interests that may not be compatible with their long term interests. If ANS actions should harm the mid-levels, the NSF would be likely to find itself in a position where it would feel it had no other choice but to step into the monetary breach.f

The NREN is seen by many as a network that will bring about major improvements in the nation's educational infrastructure. However, because the NSF does not have unlimited funds at its disposal Vice President - Elect Gore's dream of an NREN that serves a critical information age public interest purpose of enabling that little girl in Tennessee to log into the Library of Congress and do her homework on dinosaurs is in danger.^g The NREN is seen by others as technology that can be transferred to business in productive ways. According to Tom Valovic, "at least in theory, with the proper oversight by those responsible for its development, the Internet [and its NREN component] could become a worldwide data networking utility that multinational corporations, and enterprises could tap into regardless of geographical location."³ Because of policy decisions made in 1990 and not yet rectified, this outcome is also problematic.^h

The American Internet has split into two unfriendly camps: ANS and some of the mid-levels on the one hand, and the Commercial Internet exchange on the other. The continuance of network connectivity between the two camps is not assured, and the NSF admits that there will be no single NREN network.ⁱ Probably through no fault of its own the NSF has lost control.^j But OSTP, and Congress do not even seem to recognize that there is a problem.

Senator Gore's concept of an NREN was the right technology choice. But it is one that threatens to be still-born because, for the past two years, the wrong decisions have been made about its implementation. While in Japan industries are re configuring their cultures to form cooperative links both vertically and horizontally, we had in Republican America a counter-intuitive obedience to the free market with a concomitant belief that what is good for our biggest corporations is good for the nation.^k This is the policy that drove decisions in ruinous directions. Decisions were made to "privatize" the NSFnet backbone now rather than wait for the emergence of a privatized NREN that congressional legislation called for before the end of the decade.^l

In 1990 MERIT, IBM and MCI decided to create a jointly funded spin off called ANS that would privatize the NSFnet backbone. While the plan was carried out, the new venture quickly found itself in conflict with almost all other network players. More than two years of strife have followed where, instead of cooperation, the government endowed apparent favorite found itself in divisive conflict with national as well as regional service providers. Documents obtained from the National Science Foundation under the Freedom of Information Act provide the critical missing pieces to what is otherwise a part of the public record.

Change is needed if this critical technology is to serve the national interest. But to be able to change, those involved must realize that what they are doing is unlikely to succeed, that the model they are building is broken. This essay will show the brokeness of the model and will close with recommendations for change.