Bookmark and Share

Summary and Observations

The evolution of packet switching networks can be briefly summarised. The early implementations, with one exception, embodied the assumption that networks should be responsible for reliably delivering packets because an unacceptable burden would otherwise be placed on their users. ARPANET was based on this principle and so were the X.25 networks, unsurprisingly since they were inspired by ARPANET. By contrast, CIGALE and Internet (IP) did not assume responsibility for reliable delivery and were therefore dependent on software on the host computers - Station de Transport and TCP respectively. The mainframe manufacturers, having network architectures of their own, were not motivated to build TCP/IP into their offerings. The new workstations and servers, however, used UNIX which had TCP/IP built in and they were about to experience a boom as users turned away from the mainframe centred information culture to seek more flexible options.

ARPA was successful, in part, because it spotted and backed winners like UNIX and Ethernet. And its timing, intentionally or unintentionally, was impeccable. Those who bought into the new style of computing did so voluntarily because it met their needs and it was available. Just as X.25 was adopted in the 1970s when TCP/IP was incomplete, TCP/IP was adopted in the 1980s when OSI, X.25's successor, was not ready.

Finally, Internet was able to proliferate because of the global infrastructure already in place. This was facilitated by IP/X.25 gateways and supplemented by new IP-only backbones. Over time X.25 was eroded by IP.

That the pieces of the story so neatly slotted into place was not entirely the result of coincidence. The roles of large corporations' labs as innovators and of government, both as paymaster and regulator, were also crucial. The latter, in particular, should give pause for thought to those for whom government involvement is anathema, believing that free markets will produce the best solutions unaided and that funding of all research should be based on predictable benefits to business[1].

In the United States we have this belief that competition, that deregulation guarantees fair competition, and my experience is that ain't so. [...] So there is a tension, serious tension, between openness and this ability to preserve old business models (V. Cerf 2007, 22).

Cerf is referring to what he regards as the slow roll-out of broadband in the US. However, these remarks raise a more general question about the role of the Justice Department and antitrust in the developments described here. Monopolistic behaviour is not difficult to locate: IBM sitting on RISC in case it undermined its mainframe business, Xerox charging for copies by the sheet, flatly refusing to concede the feasibility of a desktop copier, and AT&T's general disinterest in developments in data communications that might undermine its voice telephony business.

The most obvious benefit of antitrust action was the liberation of UNIX from AT&T. By the mid-1990s vendors supplying UNIX servers had proliferated - Sun, HP, Cray, DEC and IBM had been joined by Amdahl, Texas Instruments, Siemens Nixdorf, Olivetti, Fujitsu, NCR, Data General and others[2].

IBM's operations had been characterised by a compulsion to dominate every market in which they competed. When they had no product to offer, their formidable marketing arm would make declarations of intent calculated to spread uncertainty and paralyse the market. Few believed that when IBM broke with tradition and announced the PC, in a form that would not permit them to control the market, that they had had a damascene conversion to openness. Various explanations have been advanced: they thought it was a home computer in a small market and didn't much care (Carroll 1994, 22-23) and they were in a hurry and couldn't mobilise internal resources quickly enough (Heller 1994, 53) (Ceruzzi 2003, 270), they didn't have the skills to develop the operating system software in-house (Campbell-Kelly and Aspray 1996, 254). There may be some truth in all of these but do they constitute the explanation?

The PC was announced in 1981, one year before the Justice Department dropped its antitrust suit. The suit "hung over the company's head like the Sword of Damocles" leading to an "internal circle-the-wagons mentality" (Wallace and Erickson 1992, 166). Could the PC have been a demonstration by IBM that it was a reformed character, committed to openness? Be that as it may, the result was a massive proliferation of "IBM PCs" from other vendors: HP, Texas Instruments, Compaq, Dell, Sony, Toshiba, Samsung, Acer and many others.

Xerox was less the victim of antitrust than of its own monopolistic tendencies. If Xerox had innovated in its copier business, it may have been possible to invest time and money in its computer business. We cannot know what would have happened. In the event, its technologies leaked out of PARC and were exploited by others such as Apple, 3Com, Novell, HP, and Microsoft.

In each case, a competitive market formed around the work from the laboratories of these large companies. Subsequently, the notion of a free market as a competitive market has been eroded by the deregulated market consensus, with monopoly regarded as preferable to government intervention. It was in the fertile soil of the former that Internet grew and prospered.

As David Edgerton points out, there is a great difference between innovation and the adoption and use of a technology (Edgerton 1999). The Internet is infrastructure, in itself of no more worth than electricity generation and distribution without lighting and appliances. The early switched networks (X.25) prospered by enabling new relationships between firms and their customers, such as point of sale transactions and electronic funds transfer. Internet was dragged along by the demand for personal and departmental computing and later by the need to easily locate and access documents.

In this respect, these networks have similarities to the growth of the electric telegraph network in the nineteenth century, which for many years had been a solution in search of a problem. The first telegraph deployment (1837) was driven by railway safety and the need for better signalling (Fahie 1884, 40). The Atlantic cable (1866) was motivated by finance and the need to connect the London and New York markets (Gordon 2002, 34). The London-Karachi cable (1864) was driven by imperial military needs following the Indian Mutiny (Harris 1969, 170).

In the case of Internet, intermediate technologies were needed to connect it to the end users: UNIX, Ethernet, the intelligent workstation and later the Web. Following a twentieth century pattern, also noted by Edgerton, these innovations came from the labs of large companies (Edgerton 2008, 184-203) and government. These technologies were not just passive ingredients in the recipe that became TCP/IP. Ethernet, in particular, helped to shape it and, having done so, was largely instrumental in its adoption because it met current business needs.

This was a complex process involving not only the adopters of TCP/IP but also government (as paymaster and regulator), universities, computer manufacturers and telecommunications companies. It is this perspective which helps explain the success of Internet and the demise of X.25.


1. The US Department of Defense had not imagined that Internet would be the outcome of their investment any more than the European governments that invested in CERN had anticipated the Web as opposed to, say, a glimpse of a passing Higgs Boson.

2. By the time of writing even IBM mainframes (System z) were running UNIX (Linux).


Campbell-Kelly, Martin, and William Aspray. Computer: A History of the Information Machine. New York: BasicBooks, 1996.

Carroll, Paul. Big Blues: The Unmaking of IBM. London: Weidenfeld & Nicolson, 1994.

Cerf, Vinton, interview by Donald Nielson. Oral History of Vinton (Vint) Cerf Mountain View, CA: Computer History Museum, (7 November 2007).

Ceruzzi, Paul E. A History of Modern Computing. Cambridge, MA: The MIT Press, 2003.

Edgerton, David. "From Innovation to Use: Ten Eclectic Theses on the Historiography of Technology." History and Technology 16, 1999: 111-136.

Fahie, J J. A History of the Electric Telegraph to the Year 1837. London: E & F N Spon, 1884.

Gordon, John Steele. A Thread Across the Ocean: The Heroic Story of the Transatlantic Cable. London: Simon & Schuster, 2002.

Harris, Christina Phelps. "The Persian Gulf Submarine Telegraph of 1864." The Geographical Journal (Blackwell Publishing) 135, no. 2 (Jun 1969): 169-190.

Heller, Robert. The Fate of IBM. London: Warner, 1994.

Wallace, James, and Jim Erickson. Hard Drive: Bill Gates and the Making of the Microsoft Empire. New York: Wiley, 1992.