The Internet Origins: ARPANET and Decentralized Networks

Core Idea

The ARPANET, developed by the Advanced Research Projects Agency beginning in 1968, was designed to enable computer communication across geographically distributed research institutions. Its architects — Larry Roberts, Bob Taylor, Vint Cerf, Bob Kahn, and others — developed packet-switching protocols that allowed data to be broken into chunks, sent independently across heterogeneous networks, and reassembled at the destination. This architecture was radically decentralized: no single control point, no requirement for a central authority, resilient to partial failures. The TCP/IP protocol suite, developed in the 1970s, provided a standardized way for different networks to interconnect. The ARPANET evolved into the Internet, and the development of the World Wide Web in 1989 made it accessible to non-specialists. The Internet's history reveals how technological design embodies political assumptions: the decentralized architecture reflected Cold War concerns about survivability of command networks, yet it also enabled democratic communication structures. The Internet is a rare case where a transformative technology's origins are well-documented and traceable to specific decisions.

Explainer

The Internet's history is among the best-documented cases of how a publicly-funded research project can produce a technology transforming every domain of human activity in ways its creators did not anticipate.

The ARPANET, established by the Department of Defense's Advanced Research Projects Agency beginning in 1968, had a specific and limited purpose: enabling researchers at different universities and government labs to share computing resources — expensive mainframe computers — without physically visiting each other's sites. The network was small: the first message was sent from UCLA to the Stanford Research Institute in October 1969. Larry Roberts (ARPA program manager), Bob Taylor (who pushed for the network's creation), and the contractor teams at Bolt Beranek and Newman built the first packet-switching network.

Packet switching — the fundamental architectural choice — was radical. Rather than establishing dedicated circuits (as in the telephone network), data was broken into packets that traveled independently through the network and were reassembled at the destination. This made the network more efficient and more resilient: packets could be rerouted if nodes were unavailable. The TCP/IP protocol suite, developed by Vint Cerf and Bob Kahn in the early 1970s, standardized how different networks communicated — enabling an 'internet of networks' (hence 'internet').

Through the 1970s and 1980s, the Internet was used primarily by researchers and academics for email, file transfer, and remote access. Acceptable use policies explicitly prohibited commercial activity. NSF took over civilian internet infrastructure from ARPA in the late 1980s and began transitioning toward commercial operation.

Tim Berners-Lee at CERN invented the World Wide Web in 1989 — a system of hypertext documents linked by URLs, accessed via a browser using HTTP. This was a particular application built on top of the Internet, not the Internet itself. The Web made the Internet accessible to non-specialists: point-and-click navigation through linked information spaces. Mosaic (1993), the first graphical browser, triggered rapid adoption; Netscape commercialized the Web and sparked the dot-com boom. The Internet's user base grew from roughly 1 million (1993) to over 100 million (2000).

The Internet's political dimensions were evident from the start. The network's end-to-end architecture — intelligence at the edges, dumb network core — embeds assumptions about decentralized control that contrasted with the telephone company's centralized architecture. These choices are not technically inevitable: authoritarian states have built internet architectures that filter and monitor content at the network core. The architecture of communication systems is a political choice.