The US government resisted the Internet, instead of supporting it

The question of "Who really invented the Internet?" has been one of the technological issues in the current US presidential election. This is odd, because the Internet's history has been well documented, and most of the people involved are still alive. However, the controversy has less to do with technology than with the endless battle between publicly-funded good works and private enterprise.

President Obama has identified the Internet as a public good produced by the US government. In a campaign speech, he said: "The Internet didn't get invented on its own. Government research created the Internet so that all companies could make money off the Internet."

The first two parts of that statement are true, and the final part isn't. The Internet (a network of networks) was created directly as a part of the research efforts of DARPA, the US government's Defense Advanced Research Projects Agency. However, it wasn't done so that anybody could make money from it. Indeed, in the net's early days, it was specifically reserved for non-commercial use, and the idea of making money from it was sometimes described as illegal.

Briefly, the Internet was made possible by Vint Cerf and Bob Kahn, when they developed TCP, the Internet protocol. Both men had worked on the development of the ARPANET network from about 1968. Also, while working for BBN, Kahn had developed the ARPANET IMP (Interface Message Processor), an early packet switch. The next step was to connect the ARPANET to other research networks such as NSFNET (the National Science Foundation network), NASA's SPAN (the space physics network), and MILNET to create a network of networks. That needed a common protocol, and they all had to be persuaded to support TCP/IP.

None of this means that the US government, or its various agencies, wanted the Internet. In fact, they tried to avoid adopting it.

The US and European Union governments adopted a different networking strategy, based on a complex seven-layer Open Systems Interconnection (OSI) model. This was developed in Europe as an international standard (ISO/IEC 7498-1), and it had a vast number of supporting standards. These included ITU (International Telecommunications Union) X standards such as X.25 packet switching and X.400 email.

US and EU government support for OSI led to the publication of GOSIP (Government Open Systems Interconnection Profiles) as a requirement for government procurements. In other words, commercial companies would be forced to support the local GOSIP before governments would buy their stuff.

This had a laudable aim. Most governments had bought vast quantities of incompatible equipment from global IT manufacturers such as IBM, DEC and Data General, as well as "national champions" such as ICL in the UK, Honeywell-Bull in France, Siemens and Nixdorf in Germany, and so on. They wanted it to work together, and setting an open standard was the best way to do it.

Companies had fair warning, and worked on their OSI interoperability from the early 1980s. However, the first US government specification requiring OSI protocols, FIPS 146-1, wasn't published until 1990.

In other words, the US government mandated OSI networking seven years after TCP/IP had been installed on its own research networks on January 1, 1983.

The US government didn't officially change its mind until 1995, when the new FIPS 146-2 profile allowed ITU, ISO and — crucially — IETF (Internet Engineering Task Force) standards. That enabled US government departments, commercial companies and individuals to unite behind Internet standards for the first time.

I reported OSI's development for many years, and was aware that the Internet Protocol (IP) had had to overcome tremendous resistance. I raised the point with Vint Cerf when I interviewed him for the Guardian in November 2000 (Surfing through space). He said:

"It's true, many people resisted it. Its predecessor, the ARPANET, was considered a silly idea that wouldn't work, and it was ridiculed by people who grew up in the telephone tradition. Most of the computer science communities also rejected the idea of connecting up, and ARPA had to insist.

"When it came time to convert from the old ARPANET protocols to the new Internet protocols, in January 1983, there was tremendous resistance. Finally, we had to force it on people by turning off the old protocols, so we jammed it down their throats. Then came the lengthy debate between the OSI and Internet protocols, and again, that was a 10-year battle."

President Obama is right to say that "The Internet didn't get invented on its own," and we all owe a huge debt to DARPA in particular. It was US government research funding and sponsorship from 1968 through 1995 that made the Internet a success.

But the idea that the US government did this knowingly and with commercial foresight, like building a new highway system, just isn't true. The networking system that the US government knowingly supported with commercial foresight was OSI, and that failed.