Above losing a space-race, the biggest fear for the world in the 1950's was the threat of a nuclear war. Dropping atomic bombs on Japan demonstrated the incredibly destructive power of nuclear weapons, and both sides had the bomb now. While some research centers worked on making weapons even more destructive, other researchers studied how to survive an atomic war. Protecting the nation's modes of communication was considered one of the most critical priorities.

Scientists at the RAND Corporation, a think-tank devoted to national defense, studied several possibilities. One of their scientists, Paul Baran, theorized that a decentralized network with several possible routes between any two points could keep the channels open for communication. If a few of the routes in the network were destroyed in a nuclear attack, messages would be rerouted automatically. In order to do this, though, he realized that the messages would need to be split up and sent as separate blocks. If the message was cut up into blocks, each could travel along any route that connected the source to the destination and at least part of the message would make it through.

In 1965, Baran found funding from the Air Force, but the project was plagued with bureaucratic problems. Baran was afraid that the project was doomed to fail because of the people put in charge of it, so he withdrew his request because he feared a failure would ruin any future prospects. He gave up on the idea, but he didn't know that other engineers were already working on the same idea.

Earlier in 1961, Leonard Kleinrock wrote his Ph.D. thesis at MIT on a similar block switching idea. Also, across the Atlantic in Britain, Donald Watts Davies was working on a block-switching scheme for the British National Physical Laboratory (NPL). However, Davies had a different name for it; he called the blocks "packets."

Flash 3 Demo of Packet Switching

Acceleration