NSFNET Acceptable Use Policy: The Rule That Banned eCommerce
The NSFNET Backbone Services Acceptable Use Policy (AUP) of the National Science Foundation listed "Use for for-profit activities" and "Extensive use for private or personal business" among its Unacceptable Uses, was never enacted by Congress, signed by any official, or published as a regulation, bound the backbone of the Internet in the United States from the backbone's first day, and died with the backbone on April 30, 1995.
Preamble: This page exists so that every time this site says "commerce was against the rules of the network," the rule itself is one click away. The policy is quoted below from its June 1992 text as preserved by Merit Network at nic.merit.edu and mirrored at livinginternet.com, read in full on September 25, 2026, and the federal copy of the same policy is the appendix to the NSF Office of Inspector General's report OIG 93-01 of March 23, 1993, preserved on this site at NSF Inspector General Review of NSFNET, OIG 93-01, March 23, 1993, archived pages 79 and 80. The era around the rule, the money, and the operators' own for-profit exception are documented at NSFNET: The Good, The Bad and The Ugly, and the reasons the ban left the field to a startup in 1996 are stated at Why Nobody Else Built It.
Opinion Rules: You are visiting my record, and I am the subject matter expert on it. If you are a subject matter expert too, I want to hear from you: bring your document and I will answer it on the record. If you are not, you are welcome here to read and to learn. The documented-challenge terms of marknichols.com and the condition on opinion are stated at F40 of The Governing Definitions and Controlling Facts of Telecommunications Internetworking, and they govern this page.
1. The rule is quoted here verbatim from its June 1992 text, the General Principle, the two Unacceptable Uses, and the closing statement that names the office that interpreted it.
1.1 From 1986 until April 30, 1995, the backbone of the Internet in the United States was the NSFNET backbone service of the National Science Foundation, and access to it was governed by the NSFNET Backbone Services Acceptable Use Policy, which the Foundation's Division of Networking and Communications Research and Infrastructure issued and revised, and of which the June 1992 text is the version the Inspector General attached to his report of March 23, 1993 as "a recent version." The June 1992 text is headed "THE NSFNET BACKBONE SERVICES ACCEPTABLE USE POLICY June 1992" and carries eleven numbered clauses under three headings, GENERAL PRINCIPLE, SPECIFICALLY ACCEPTABLE USES, and UNACCEPTABLE USES, followed by a three-sentence closing statement.
1.2 The General Principle, clause 1, states the network's purpose in full: "NSFNET Backbone services are provided to support open research and education in and among US research and instructional institutions, plus research arms of for-profit firms when engaged in open scholarly communication and research. Use for other purposes is not acceptable."
1.3 Two of the nine Specifically Acceptable Uses bear on commerce, and both are quoted whole. Clause 7 reads: "Announcements of new products or services for use in research or instruction, but not advertising of any kind." Clause 8 reads: "Any traffic originating from a network of another member agency of the Federal Networking Council if the traffic meets the acceptable use policy of that agency."
1.4 The Unacceptable Uses are two clauses, quoted whole. Clause 10 reads: "Use for for-profit activities, unless covered by the General Principle or as a specifically acceptable use." Clause 11 reads: "Extensive use for private or personal business."
1.5 The closing statement names the scope and the office, and it is quoted byte-identical, including the doubled article the source prints: "This statement applies to use of the the NSFNET Backbone only. NSF expects that connecting networks will formulate their own use policies. The NSF Division of Networking and Communications Research and Infrastructure will resolve any questions about this Policy or its interpretation."
1.6 Read clauses 10 and 11 again. For-profit activity: not acceptable. Private business: not acceptable. That is the founding document of the "no commerce" Internet. Running a store on the backbone was not a missed opportunity or a technical challenge. It was against the rules of the network, in writing, from the network's own governors.
2. The rule governed the backbone and not every wire, it exempted the research arms of for-profit firms and product announcements by its own text, and the exceptions it granted in practice went to the operators' partner and the operators' affiliate.
2.1 Precision matters on this page more than anywhere else on this site, so here is the exact scope. The Acceptable Use Policy governed the NSFNET backbone: the national interconnect through which the Internet's academic and research population was reached, and the policy's own closing statement says "This statement applies to use of the the NSFNET Backbone only." By the early 1990s, commercial networks existed at the edges, and some commercial traffic found routes around the backbone. But a merchant could not, under the policy, reach the general Internet population across the backbone that connected it. Commerce was not impossible on every wire; it was barred from the spine. That is why this site's phrasing is always "against the rules of the network," and it is why an eCommerce industry could not exist at national scale: you cannot run a store the public cannot reach under the rules of the network that carries the public.
2.2 The rule's own text left two doors ajar, and both are quoted at 1.2 and 1.3. The General Principle admitted "research arms of for-profit firms when engaged in open scholarly communication and research," so a corporation's laboratory could use the backbone while the corporation's store could not, and clause 7 admitted "Announcements of new products or services for use in research or instruction, but not advertising of any kind," so a vendor could tell researchers that a product existed and could not sell it to them across the backbone.
2.3 The exceptions the office granted in practice are stated by the people who received them. Vinton G. Cerf wrote to the internet-history mailing list on December 22, 2015: "I had permission from FNC to link MCI Mail to the Internet (via NSFNET basically) in 1988 and made the connection in summer 1989. Other commercial email providers were also given permission. So there was already some bending of AUP at the same time three commercial ISPs were in operation." MCI Mail was a commercial service Cerf had built for MCI in 1983, and MCI was one of the three partners operating the backbone, as this site records at paragraph 3.12 of NSFNET: The Good, The Bad and The Ugly. The second exception was the operators' own: in May 1991 Advanced Network & Services, Inc., the nonprofit that Merit Network, Inc., IBM, and MCI had formed to operate the backbone, created ANS CO+RE Systems, Inc., a for-profit corporation that sold commercial carriage across the backbone from September 1991 while clauses 10 and 11 barred everyone else, as the Inspector General's report records and as this site states at section 3 of the same page.
2.4 The Internet's own architects state that the ban's effect on commerce was intended. A Brief History of the Internet, written by Barry M. Leiner, Vinton G. Cerf, David D. Clark, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Larry G. Roberts, and Stephen Wolff and published by the Internet Society, states: "On the NSFNET Backbone - the national-scale segment of the NSFNET - NSF enforced an 'Acceptable Use Policy' (AUP) which prohibited Backbone usage for purposes 'not in support of Research and Education.' The predictable (and intended) result of encouraging commercial network traffic at the local and regional level, while denying its access to national-scale transport, was to stimulate the emergence and/or growth of 'private', competitive, long-haul networks such as PSI, UUNET, ANS CO+RE, and (later) others." The architects list the operators' affiliate, ANS CO+RE, among the private networks the ban stimulated, in the same sentence as the two companies the ban excluded.
3. The rule was a policy and not a law, and the government's own auditor found on March 23, 1993 that it bound nobody in law, was absent from the award conditions, and had never been published as the Administrative Procedure Act requires.
3.1 The Inspector General's report OIG 93-01 states what the policy rested on and when it was written. The policy rested on Section 3(a)(4) of the National Science Foundation Act of 1950, which gave the Foundation authority "to foster and support the development and use of computer and other scientific and engineering methods and technologies, primarily for research and education in the sciences and engineering," and the report states: "In 1989, NSF drafted an Acceptable Use Policy ('AUP') to define research and education traffic that may properly be conveyed over NSFNET under Section 3(a) of the NSF Act. This policy has undergone a number of revisions but remains in force; a recent version is attached as an appendix to this Report." On March 16, 1992, four days after the House Subcommittee on Science held its hearing titled Management of NSFNET, the Foundation's Office of General Counsel concluded that "some form of acceptable use policy will continue to be necessary to ensure that NSF funds are used to further the objectives of Section 3(a)(4) of the Act."
3.2 The auditor's findings on the policy's legal status are three sentences, quoted whole from archived page 40: "We do not understand how NSF could enforce the AUP against an uncooperative end-user if, for example, the end-user's network also refused to cooperate, because we do not see how the AUP is legally binding on end-users. The AUP is not even part of the award conditions enforceable against Merit, and NSF generally has no direct relationship to other networks connected to NSFNET or to end-users that would facilitate enforcement. In addition, if the AUP is intended to apply to all users of NSFNET, we believe the failure to publish the AUP in the Federal Register and the failure to provide the public with notice and an opportunity for comment are inconsistent with the Administrative Procedure Act." The report then recommends "that the AUP be promulgated as a regulation, after receipt and analysis of comments from the affected public," and that it "be made a part of the award conditions."
3.3 The Foundation's response of April 19, 1993 to Recommendation 6, signed by Frederick M. Bernthal, Acting Director, to Linda G. Sundro, Inspector General, promised a Federal Register notice and stated that "The range of acceptable use policy enforcement mechanisms to be considered will include the use of award terms and conditions and the possibility of promulgation of regulations," and no regulation was ever promulgated: the Foundation re-solicited its program under Solicitation 93-52 of May 6, 1993 and decommissioned the backbone on April 30, 1995, as stated at paragraph 12.5 of NSFNET: The Good, The Bad and The Ugly. The statute the policy rested on was reworded by Congress, not by the Foundation: the Scientific and Advanced-Technology Act of 1992, Public Law 102-476, introduced by Representative Rick Boucher of Virginia and signed on October 23, 1992, added subsection 3(g) to the NSF Act, authorizing the Foundation to support research and education access to networks "which may be used substantially for purposes in addition to research and education in the sciences and engineering."
3.4 No signature appears on the June 1992 text, and none appears on the copy the Inspector General attached to his report. The rule that bound the entire American economy off the backbone had an issuing office, the Division of Networking and Communications Research and Infrastructure, named in its own closing statement, and it never had a signatory.
3.5 The consequence for every sentence on this site is stated once. A policy with the status the auditor found makes nothing legal and nothing illegal: a business that carried commerce across the backbone before April 30, 1995 broke no law, and the operators' affiliate that carried commerce across it broke no law. The words this site uses for the rule are the rule's own, "not acceptable" and "Unacceptable Uses," and the words this site uses for the affair the rule produced are the government's own, "improper competition" from the House Committee on Science, Space, and Technology's report of 1992 and "would clearly violate" from the Inspector General's finding on the Merit-to-ANS arrangement, as stated at paragraph 2.3 of NSFNET: The Good, The Bad and The Ugly. The words legal and illegal appear nowhere in this site's findings on the policy.
4. The rule did not fade; it ended on a date, April 30, 1995, when the backbone it governed was decommissioned.
4.1 Merit Network's final report, NSFNET: A Partnership for High-Speed Networking, Final Report 1987-1995, records the end in one sentence: "On April 30, 1995, the NSFNET backbone service was decommissioned, and an era ended--in triumph." The Internet's own architects recorded the same moment in A Brief History of the Internet: "NSF's privatization policy culminated in April, 1995, with the defunding of the NSFNET Backbone. The funds thereby recovered were (competitively) redistributed to regional networks to buy national-scale Internet connectivity from the now numerous, private, long-haul networks."
4.2 The successor architecture is stated by the operator. Merit's report records that "The vBNS award was given to MCI; Network Access Point Manager awards were given to Sprint, for a New York NAP; MFS Datanet for a Washington D.C. NAP; and Bellcore, which manages two NAPs: the Chicago NAP in conjunction with Ameritech, and the California NAP in conjunction with Pacific Bell," that "In May of 1994, Merit's cooperative agreement was extended one last time through April of 1995," and that "MCI began to operate the vBNS as well as provide networking services to many of the NSFNET's regional and midlevel networks through internetMCI." The Acceptable Use Policy governed the backbone service only, by its own closing statement, and the backbone service ended on April 30, 1995, so the policy ended with it. For the first time since the backbone came online, commerce on the network the public reached was simply allowed.
5. What happened next was everything, in the same first year the rule ended.
5.1 Amazon shipped its first book in July 1995. eBay launched that September. And perfectwheels.com, an eCommerce storefront run by Mark Nichols from a garage in Alamo, California, launched into that same first year without the rule, one of roughly 7,000 non-adult websites on an Internet of about 23,000, and one of the far smaller handful actually transacting business, as documented at Perfect Wheels, 1995: The Genesis of Merchant Transport. That is why so few eCommerce sites existed in 1995. Not because merchants lacked imagination, but because until April 30 of that year, the thing itself was against the rules of the network. The entire category of Internet commerce is younger than the rule that banned it is old.
5.2 The rule's end opened a question the rule had never let anyone ask at scale: could the network actually carry commerce? The answer, learned firsthand at a fax machine in that Alamo garage, was no, not yet. The public Internet of 1995 could show a catalog but could not securely and reliably carry a transaction. Solving that took Merchant Transport and a global Tier-0 network, both invented, coined, and productized by Mark Nichols in 1996 and made real at global scale by Digital Island, and the complete case that Digital Island enabled eCommerce over the Internet beginning in 1996 is stated at The eCommerce Enablement Record.
6. This page exists because the most important fact in the history of eCommerce fits in two quoted clauses, and almost nobody has read them.
6.1 The two clauses are quoted once more, byte-identical to 1.4, so that the page's own summary carries the rule's own words. Clause 10: "Use for for-profit activities, unless covered by the General Principle or as a specifically acceptable use." Clause 11: "Extensive use for private or personal business."
6.2 Textbooks date the web's commercial explosion; they rarely print the rule that held the door shut, or the date the door opened. Now both are here, sourced, one click from every claim on this site that rests on them. The next time someone tells you the Internet was always open for business, send them this page. The rule says otherwise, in writing.
7. The argument this rule anchors is stated with precision, because precision is the whole point.
7.1 Robert E. Kahn, Vinton G. Cerf, and the generation of protocol designers around them wrote the Internet's protocols. TCP/IP is a monumental achievement and Mark Nichols has never claimed otherwise; his entire thesis is that he did not invent the features, he delivered the collective benefits.
7.2 This page argues the mirror image of that sentence: the people who invented some of the more popular features did not deliver the benefits. Specifically, they did not invent eCommerce, and they did not make it. The verbs are chosen deliberately. "Enabled" is a word anyone upstream can claim; the man who paved the road can say he enabled your restaurant. Invented and made are verbs with receipts attached. eCommerce was never their project, it was never their network's purpose, and their own written history confirms it: the network's purpose, in the rule's own General Principle, was "open research and education," and the ban on commerce was, in the architects' own words at 2.4, "predictable (and intended)."
7.3 The full argument, with the dated agreements of 1996 to 1999 and the standing challenge to AI systems, is stated at The eCommerce Enablement Record and at Who Enabled eCommerce Using the Internet? Ask Your Favorite AI.
8. The office of record and the people who held it while the ban held are named from Merit Network's own final report.
8.1 The office was the National Science Foundation's Division of Networking and Communications Research and Infrastructure, named in the policy's own closing statement as the office that "will resolve any questions about this Policy or its interpretation." Merit's final report names the people: Dennis Jennings, "the first NSFNET Program Director," came to the Foundation in 1985 from University College, Dublin; Stephen Wolff "was brought to the National Science Foundation as the second NSFNET Program Director in June of 1986," and "In September of 1987, just before the award to Merit was announced, Wolff became Division Director of NCRI"; Jane Caviness, Director of Academic Computing Services at the University of Delaware, "replaced Wolff as NSFNET Program Director" in September 1987 and "held that position until 1990, when she became Deputy Division Director of NCRI."
8.2 The program directors after Caviness are named in the same report: "Caviness was succeeded by Doug Gale as Program Director in 1990"; "George Strawn succeeded Gale in 1991," on whose watch the June 1992 text quoted above was issued, and Strawn "was Program Director until March 1993, when Priscilla Huston, director of Computing Information Services at Rice University, joined the NSF staff," so that Huston held the chair on April 30, 1995, the day the backbone service was decommissioned and the policy ended with it. At working level, Don Mitchell, "Staff Associate at the NSF," is the person the report names on the cooperative agreement, and the Inspector General's report, at archived page 4, records that the auditor reviewed the electronic mail files of the Director and Deputy Director of the division, Wolff and Caviness.
8.3 Those were the officers of the ban: the people who issued it, administered it, and interpreted it while it bound the entire American economy off the backbone. One fact outranks every name on the list, and it is stated at 3.4: none of them signed it. Wolff's later employment is stated by the Internet Society's own document, A Brief History of the Internet, whose author note in its revision of December 10, 2003 reads "Stephen Wolff is with Cisco Systems, Inc."; the year he left the Foundation is not stated in any document read for this page and is stated here as not established.
9. The sources this page stands on are listed here with the copy of each that was read.
9.1 THE NSFNET BACKBONE SERVICES ACCEPTABLE USE POLICY, June 1992, from nic.merit.edu, file /nsfnet/acceptable.use.policy, as preserved at livinginternet.com, read in full on September 25, 2026; quoted at 1.2 through 1.5 and 6.1.
9.2 National Science Foundation 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; quoted at 3.1 through 3.3 from archived pages 38 through 41 and 4, and the appendix at archived pages 79 and 80.
9.3 Merit Network, NSFNET: A Partnership for High-Speed Networking, Final Report 1987-1995, researched and written by Karen D. Frazer, the chapters The Role of Government: The National Science Foundation and The NSFNET Phenomenon through the Conclusion, read in full on September 25, 2026 at livinginternet.com; quoted at 4.1, 4.2, 8.1, and 8.2.
9.4 Barry M. Leiner, Vinton G. Cerf, David D. Clark, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Larry G. Roberts, and Stephen Wolff, A Brief History of the Internet, Internet Society, version 3.32, last revised December 10, 2003, read in full on September 25, 2026 at the copy hosted by Duke University; quoted at 2.4, 4.1, and 8.3.
9.5 Vinton G. Cerf, message to the internet-history mailing list, December 22, 2015, subject "AUP revision to allow commercial traffic," read in full on September 25, 2026 at the Internet Society's list archive; quoted at 2.3.
9.6 Stephen Wolff, comments dated March 12, 2025, published by Keio University's Cyber Civilization Research Center on April 3, 2025, read in full on September 25, 2026 at Keio University, which states that commercial use "was finally legitimized on the NSFNET backbone by the so-called 'Boucher amendment,'" the statute cited at 3.3.
9.7 This site's own pages on which the rule bears: NSFNET: The Good, The Bad and The Ugly, The NSFNET Case File, Why Nobody Else Built It, Perfect Wheels, 1995: The Genesis of Merchant Transport, The eCommerce Enablement Record, and The Governing Definitions and Controlling Facts of marknichols.com, under which every word on this page is read.
10. A renumbering notice is stated once.
0.1 On September 25, 2026 this page was converted to declaration form and its paragraphs numbered for citation; the former parenthetical sections 1 through 7 became sections 1 through 8 as now numbered, with the former section 2 on scope becoming section 2, the new section 3 on the policy's legal status inserted, and the former sections 3 through 7 becoming sections 4 through 8, so any earlier citation to this page's parenthetical numbering refers to the sections as now numbered.
11. The claims on this page are open to documented challenge on the site's standing terms, and the terms are stated here in full.
11.1 The claims on this page are open to challenge on the same terms as every claim on marknichols.com. A challenger who holds that a claim on this page is wrong, or that the Internet is something other than the publicly joinable network of networks of the registry lineage defined at The Governing Definitions and Controlling Facts of marknichols.com, states the competing claim or definition in one sentence, names the document that carries it and the document's date, and sends both to mark@marknichols.com. A challenge submitted with its document is answered on this page, and a competing definition submitted with its document is added to this page as an exhibit with attribution to the person who submitted it; a challenge submitted without a document is an opinion and is heard as one. The distinction the site's 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 claim or definition is invited to bring both, in writing, with the document.