ARPANET and the First Network of Distant Computers
ARPANET linked research computers through packet-switching nodes, created a community for network protocols, and provided the experimental environment from which Internet architecture emerged.
ARPANET was designed to connect incompatible research computers
In the 1960s, advanced computing resources were scarce and distributed among universities and laboratories. ARPA’s networking program sought a way for researchers at one site to use computing resources at another without requiring every institution to own the same machine.
The resulting ARPANET used packet switching and dedicated network computers called Interface Message Processors, or IMPs, to separate host computers from much of the communications machinery.[1]
The first four nodes came online in 1969
DARPA lists the original sites as UCLA, Stanford Research Institute, the University of California Santa Barbara and the University of Utah.[1] UCLA received the first IMP in August 1969, and SRI received the second shortly afterward.[2]
These institutions brought different computing systems onto one experimental network, making interoperability an immediate engineering requirement.
BBN built the Interface Message Processors
Bolt Beranek and Newman built the IMPs that formed the packet-switching subnetwork. Host computers connected locally to an IMP, while the IMPs forwarded traffic across leased links.
A separate packet network reduced host-specific complexity
The IMP architecture meant the network could evolve without requiring every research host to implement the full switching function itself. It was an early form of separating edge computers from networking infrastructure.
The first host-to-host message was only two letters long
On October 29, 1969, Charley Kline at UCLA attempted to log into an SRI computer while Bill Duvall monitored the receiving side. UCLA professor Leonard Kleinrock’s preserved account records that the letters L and O arrived before the system crashed; the full LOGIN succeeded later.[2][3]
The famous “LO” story is memorable, but the deeper milestone was successful interactive communication between two remote host computers over a packet network.
Network protocols became collaborative public documents
ARPANET researchers needed common rules for host communication. The Network Working Group developed protocols and documented ideas through Requests for Comments, creating a culture in which protocol design could be circulated, criticized and revised.
That process helped normalize open technical documentation as part of network governance.
NCP provided early host-to-host communication
By 1970 the Network Working Group had completed the Network Control Protocol, giving ARPANET applications a common host communication layer.[4]
Applications made the network worth using
Remote login, file transfer and later electronic mail turned ARPANET from an infrastructure experiment into a daily communications environment for its community.
The network expanded beyond its four-node beginning
DARPA’s historical map shows ARPANET spreading across universities, government laboratories and military research sites during the 1970s.[1] International links also connected the network to sites outside the United States.
Growth made the weaknesses of a single-network architecture increasingly visible.
ARPANET exposed the difference between networking and internetworking
A single packet network could connect many hosts, but DARPA soon had packet radio and satellite networks with different technical characteristics. The next challenge was to connect networks to one another without forcing them to become identical.
That requirement led Bob Kahn and Vint Cerf toward the protocol architecture that became TCP/IP.[5]
The 1983 TCP/IP transition changed ARPANET’s role
On January 1, 1983, ARPANET hosts transitioned from NCP to TCP/IP. The Internet Society describes this as a carefully planned flag-day conversion that enabled a broader network-of-networks model.[5]
ARPANET increasingly became one component inside the larger Internet rather than the Internet’s complete identity.
Why ARPANET matters without calling it ‘the Internet’ too early
ARPANET was a crucial predecessor and experimental platform, but the Internet is defined by internetworking across multiple networks. Distinguishing the two preserves the history of both packet networking and the later TCP/IP architecture.
ARPANET’s central achievement was creating a working community in which heterogeneous computers, protocols, network applications and open technical coordination could develop together.
Works Cited
- 01DARPA — ARPANET darpa.mil
- 02Leonard Kleinrock / UCLA — An Early History of the Internet lk.cs.ucla.edu
- 03UCLA — Internet 50 Anniversary and First ARPANET Message newsroom.ucla.edu
- 04Computer History Museum — Internet History of the 1970s computerhistory.org
- 05Internet Society — A Brief History of the Internet internetsociety.org
CodeHistory is a living archive. Citations document the evidence used for this edition; later evidence may refine the account.
Submit a research lead