The NSFNET Case File

The NSFNET Case File is a classroom inquiry into the National Science Foundation's administration of the NSFNET backbone from 1987 to 1995, built entirely on the government's own documents, in which the respondent is the institution, the persons named are the evidence of what the institution did, and the verdict is labeled as the exercise's finding and never as the record's.

Preamble: This page is a teaching instrument and states so in every part. Section 1 names the respondent and the standards the government wrote for itself; section 2 states ten counts, each a sentence of the government's own words with a citation into the record page; section 3 lists the persons in evidence with the document that places each in the record; section 4 lists the exhibits; section 5 states the witness the record lacks; and section 6 states the form of the exercise and its verdict. No person is charged with anything on this page, the government's own auditor found no violation of law and made no referral, and the exercise adjudicates a policy record against the standard the government set for itself. The record this page cites is stated in full at NSFNET: The Good, The Bad and The Ugly, and the Inspector General's report is preserved verbatim at NSF Inspector General Review of NSFNET, OIG 93-01, March 23, 1993.

1. The respondent is the National Science Foundation, and the standards it is measured against are five documents the government wrote before or during the conduct.

1.1 The respondent is the National Science Foundation, which awarded Merit Network, Inc. a five-year cooperative agreement in November 1987 to operate the NSFNET backbone, which paid $57.9 million for the backbone service over seven and one-half years by Merit's own final report, and which decommissioned the backbone on April 30, 1995, as stated at NSFNET: The Good, The Bad and The Ugly 1.1, 3.19, and 7.1.

1.2 The first standard is NSF Important Notice 91 of March 11, 1983, quoted by the Inspector General at archived page 33, forbidding NSF awardees to use NSF-supported facilities commercially in a manner that "may have a material and deleterious effect on the success of private companies engaged in the provision of equivalent services."

1.3 The second standard is NSF Manual 10, quoted by the Inspector General at archived pages 33 and 34, requiring National Science Board approval for "a policy issue that has not previously been resolved by the Board" and for "a new program or novel set of circumstances."

1.4 The third standard is Special Condition E.3 of the Cooperative Agreement, quoted by the Inspector General at archived page 46, requiring that "any agreements between the awardee and other participants for the purpose of activities of mutual benefit must receive the NSF Grants and Contracts Officer's written approval" before execution.

1.5 The fourth standard is Attachment O to Office of Management and Budget Circular A-110, quoted by the Inspector General at archived page 50, forbidding participation in federally funded contracting by any organization that "has a financial interest" and requiring "open and free competition."

1.6 The fifth standard is the Administrative Procedure Act, cited by the Inspector General at archived pages 40 and 41, requiring publication of "statements of general policy or interpretations of general applicability," which the Acceptable Use Policy never received.

2. The counts are ten sentences in the auditor's own words, each cited to the record.

2.1 Count one, authorization without reasoning. "The record is utterly barren of documentation of NSF's reasoning for allowing commercial use of the network," and the only documentation was "an electronic mail message and a letter, both from the Director of DNCRI," Stephen S. Wolff, dated September 10, 1990 and May 24, 1991, as stated at NSFNET: The Good, The Bad and The Ugly 5.1, 11.2, and 11.3, from archived pages 21 through 23 and 31.

2.2 Count two, bypass of the Board. "The decision to (3) allow commercial use of the T3 network was not reviewed by the Board," although it was "a policy issue that ha[d] not previously been resolved by the Board," as stated at 3.8 and 5.1, from archived page 34.

2.3 Count three, assignment without prior approval. "There is no record in DGC's files of DGC approval having been sought or obtained for anything having to do with the assignment to ANS; nor is there anything in the program files," and the Division of Grants and Contracts wrote on November 17, 1992 that "a substantive portion of the effort has been subcontracted without official NSF approval," from archived pages 46 and 47; the Grants and Contracts Officer is not named in the record.

2.4 Count four, a conflicted and uncompeted arrangement. "The Merit-ANS arrangement would clearly violate those provisions of Attachment O, were they applicable: Merit clearly had an interest in ANS, and there was no effort whatever to seek competition," as stated at 11.9 and 13.6, from archived page 51, with the interest being one board seat each for Merit, IBM, and MCI, from archived page 28.

2.5 Count five, an unlawful rule enforced against the public and lifted for the affiliate. The Acceptable Use Policy was "not even part of the award conditions enforceable against Merit" and not "legally binding on end-users," from archived page 40, while ANS CO+RE Systems, Inc., created in May 1991, sold commercial carriage across the backbone from September 1991, from archived pages 22 and 35, as stated at 2.3, 12.4, and 13.2.

2.6 Count six, secrecy by default. The commercial-use agreements were not in NSF's files, a competitor obtained them by written request on October 3, 1991 and received them on October 31, 1991, published them on com-priv on December 7, 1991 as "Secret NSF-Merit/ANS Agreements," and the auditor found NSF "should have affirmatively announced this development to the networking community," from archived pages 36 and 37, as stated at 3.11 and 13.2.

2.7 Count seven, money without accounting. The only audit of Merit "did not include CO+RE and the infrastructure pool," from archived page 25, and on April 19, 1993 the Acting Director of the National Science Foundation, Frederick M. Bernthal, wrote to the Inspector General, Linda G. Sundro, that "NSF will activate its membership in the Resource Allocation Committee (the Committee that oversees the distribution of the infrastructure pool)," from archived page 81, as stated at 11.6; no record of the cost and compliance audit NSF agreed to has been found.

2.8 Count eight, files that could not be produced. "No monthly performance reports were included, most of the quarterly reports were missing, and some of the annual reports were omitted from the files," and "This sort of disorganization of program files is simply unacceptable," from archived pages 49 and 50, as stated at 3.9, 3.10, and 5.1.

2.9 Count nine, an admitted error in the award's terms. NSF's response of April 19, 1993 "acknowledges that it erred in including FDP Project terms and conditions in the Merit cooperative agreement," from archived page 83, and "No one who was responsible for the amendment is still employed by the Division of Grants and Contracts," from archived page 45.

2.10 Count ten, the disposition of the asset. The affiliate the institution permitted was sold to America Online for $35 million under an Asset Purchase Agreement dated November 25, 1994 and resold to WorldCom, Inc. on January 31, 1998 for $175 million in cash plus the online services business of CompuServe Corporation, in a transaction America Online valued at approximately $425 million, with no payment to the National Science Foundation or to the Treasury in any filing, as stated at 3.15 and 3.18 from America Online's filings cited at 10.11 through 10.13.

3. The persons in evidence are listed with the document that places each in the record.

3.1 Stephen S. Wolff, Director of NSF's Division of Networking and Communications Research and Infrastructure from September 1987, by Merit's final report: author, by office, of the letter of September 10, 1990 and the message of May 24, 1991, by the Inspector General's report at archived pages 31 and 47; witness for NSF at the House Subcommittee on Science hearing of March 12, 1992; author of the note of March 12, 2025 stating that commercial use was "finally legitimized on the NSFNET backbone by the so-called 'Boucher amendment.'"

3.2 Jane Caviness, NSFNET Program Director from 1987 to 1990 and Deputy Division Director of DNCRI from 1990, by Merit's final report; the Deputy Director of DNCRI was interviewed by the auditor and her electronic mail files were reviewed, by the Inspector General's report at archived pages 3 and 4.

3.3 A. Nico Habermann, witness for the National Science Foundation at the hearing of March 12, 1992, by the hearing record as cited at NSFNET: The Good, The Bad and The Ugly 4.1.

3.4 Frederick M. Bernthal, Acting Director of the National Science Foundation: signer of NSF's response of April 19, 1993, archived pages 81 through 86.

3.5 Linda G. Sundro, Inspector General of the National Science Foundation: issuer of report OIG 93-01 of March 23, 1993 and addressee of NSF's response.

3.6 Rick Boucher, Representative from Virginia and Chairman of the House Subcommittee on Science: convener of the hearing of March 12, 1992, author of its equal-treatment principle, requester of the Inspector General's review by the report's own first page, and introducer of the Scientific and Advanced-Technology Act of 1992, signed October 23, 1992, by archived page 38.

3.7 Douglas E. Van Houweling, Chairman of the Board of Merit Network, Inc., Vice Provost for Information Technology at the University of Michigan, and Merit's director on the board of Advanced Network & Services, Inc., by Merit's final report: witness at the hearing of March 12, 1992; quoted in Merit's final report, "In retrospect, I think we should have paid more attention to better communication about our activities and objectives."

3.8 Allan Weis, President of Advanced Network & Services, Inc., formerly of IBM Research, by Merit's final report: quoted there on the purpose of ANS CO+RE Systems, Inc., "so that if we did anything that was commercial and taxable, we could pay tax on it."

3.9 Dick Liebhaber of MCI and John Armstrong of IBM: present for their companies at the September 1990 announcement of ANS, by the caption in Merit's final report, which is the only document read that places them, and which gives Liebhaber's title as a senior vice president of MCI and gives Armstrong no title.

3.10 Vinton G. Cerf: by his own statement of July 21, 2015, asked "Around 1988" for permission from the federal interagency council he names as the Federal Networking Council to connect MCI Mail, "a commercial service I had built for MCI in 1983," to NSFNET, and the gateway ran in 1989, as stated at NSFNET: The Good, The Bad and The Ugly 3.12.

3.11 Mitchell Kapor of the Commercial Internet Exchange Association and William L. Schrader of Performance Systems International: witnesses at the hearing of March 12, 1992, by the hearing record as cited at 4.1; Performance Systems International and AlterNet are the companies Merit's final report names as having "charged that the NSF was unfairly competing with them."

3.12 The persons the record leaves unnamed: the NSF Grants and Contracts Officer whose written approval Special Condition E.3 required, referred to as "her" in Amendment 11 of November 17, 1992; the competitor who obtained and published the agreements between October 3 and December 7, 1991; and the Assistant Director for Computer and Information Science and Engineering, the Director of the National Science Foundation, and the Chairman of the National Science Board in 1990 and 1991, whom no document read places at the decision.

4. The exhibits are the documents this page is built from, with their addresses.

4.1 NSF Office of Inspector General, Review of NSFNET, report OIG 93-01, March 23, 1993, with NSF's response of April 19, 1993 and the Acceptable Use Policy as its appendix, preserved verbatim with archived page numbers at NSF Inspector General Review of NSFNET, OIG 93-01, March 23, 1993.

4.2 Merit Network, NSFNET: A Partnership for High-Speed Networking, Final Report 1987-1995, at merit.edu.

4.3 Management of NSFNET, hearing before the Subcommittee on Science, Committee on Science, Space, and Technology, U.S. House of Representatives, March 12, 1992, 190 pages, ERIC ED350986, as cited at NSFNET: The Good, The Bad and The Ugly 10.2.

4.4 America Online, Inc., Form 10-K for the fiscal year ended June 30, 1996; Form 8-K filed September 15, 1997 with the press release of September 8, 1997; and Form 10-Q for the quarter ended December 31, 1997, at the Securities and Exchange Commission addresses cited at NSFNET: The Good, The Bad and The Ugly 10.11 through 10.13.

4.5 Vinton G. Cerf's statement of July 21, 2015 as reported by Katherine Connor of The Daily Transcript on July 22, 2015, cited at NSFNET: The Good, The Bad and The Ugly 10.14; Stephen S. Wolff's note of March 12, 2025, cited at 12.7; and Jay P. Kesan and Rajiv C. Shah, Washington University Law Quarterly, volume 79, 2001, cited at 11.10.

5. The witness the record lacks is the customer, and that testimony exists in executed documents.

5.1 Every exhibit in section 4 was written by the government, the operator, the buyer, or a scholar, and none speaks for a business that sought to carry commerce across the backbone before April 30, 1995 and could not, or for what was built when the door opened. That testimony is the Digital Island record from 1996 to 2001, in executed documents, at marknichols.com, and its witness is Mark Nichols, co-founder and chief architect of Digital Island, whose expert statement in deposition form takes this page's section 7 when written.

6. The exercise is stated here, and its verdict is the exercise's finding on the evidence, labeled as such.

6.1 The exercise assigns a prosecution, a defense, an expert witness for the customer side, and a panel. The prosecution proves the ten counts from the exhibits and nothing else. The defense holds, correctly, that no law was broken, that the auditor found the decision "not unreasonable," and that the government's own auditor closed the file, and then argues whether conduct that broke no law breached the five standards of section 1. The panel decides two questions: whether the National Science Foundation breached the standards it wrote for itself, and whether an institution whose network excluded the paying public by rule from 1986 to April 30, 1995 can be credited with creating the public's Internet. The second question is decided under the definitions that govern every page of this site, The Governing Definitions and Controlling Facts of marknichols.com, under which the Internet, capital I, is the publicly joinable network of networks of the registry lineage, an internet, lowercase, is any interconnection of networks, and an intranet is a closed network, so that the panel measures the NSFNET backbone against the definition of the thing it is said to have created and not against the word alone.

6.2 The verdict form, for the panel to complete on the evidence: on the first question, the panel finds that the National Science Foundation did or did not breach its own standards, count by count; on the second question, the panel finds that the Internet, the publicly joinable network of networks, was born in 1991 at the Commercial Internet eXchange outside the NSFNET backbone and was built into the Modern Internet in the open market after February 8, 1996, as the dated record states at The Birth of the Internet, and that the credit for the public's Internet rests with the companies and capital that built it, and not with the institution whose rule kept the public out.

6.3 The sentence that closes the exercise is the claimants' own words beside their own admissions: Merit's final report calls NSFNET "the foundation of the U.S. Internet and the main catalyst for the explosion in computer networking around the world that followed," and the same report records that paying customers "were restricted from using the NSFNET backbone due to the NSF's Acceptable Use Policy"; the panel reads both, dated, and rules.

6.4 The definitions stated at 6.1 are open to challenge on the same terms as every claim on this site. A challenger who holds that the Internet is something other than the publicly joinable network of networks of the registry lineage states the competing definition in one sentence, names the document that carries that definition and the document's date, and sends both to mark@marknichols.com; the panel then decides the second question of 6.1 under both definitions, the site's and the challenger's, and the exercise records which definition the exhibits of section 4 support. A definition submitted with its document is added to section 4 as an exhibit with attribution to the person who submitted it, and a definition submitted without a document is an opinion and is heard as one. The same terms apply to the petition at Stanford University Is Petitioned to Retract the Headline of Its BIRTH OF THE INTERNET Plaque and to Keep Every Name in Its Body, whose headline question turns on the same definition, and a challenge lodged against one is a challenge lodged against both. The distinction the definition rests on, that protocols enable interoperability and are not the Internet, is stated in the IETF's own words, RFC 2026 and STD 1, in the Internet-Draft draft-mnichols-protocols-not-the-internet-01, posted by Mark Nichols to the IETF Datatracker on September 25, 2026 at https://datatracker.ietf.org/doc/draft-mnichols-protocols-not-the-internet/ and stated on this site at Protocols Are Not the Internet: The IETF Internet-Draft by Mark Nichols, so a challenger may also address the definition in the venue that publishes the protocol standards themselves. Any person who holds expertise in telecommunications and a competing definition is invited to bring both, in writing, with the document, and the exercise will run.