Better Dead Than Red: The Theory and Motivations for the InternetFrom the end of WWII through to the 1980s, fear of the Soviet Union was a powerful force that motivated nations to put more money into weapons and technology research. The United States, considered the U.S.S.R.’s primary opponent in the Cold War, was shocked when the U.S.S.R. launched a satellite called Sputnik on October 4, 1957. It was the first satellite ever to be put into orbit, and it heralded the start of a space race that would see both animals and men launched beyond gravity’s pull.
ARPA Founded to Avoid Future SputniksIn 1958, President Eisenhower created the Advanced Research Projects Agency (ARPA) in response to the Sputnik launch. ARPA was a place for the Department of Defense to focus on research and development projects so that the U.S. could avoid the unpleasant surprise of another Sputnik – that is, another nation demonstrating technological capabilities that the U.S. could not match.
One of the more terrifying aspects of Sputnik’s launch was the realization that Soviet rocketry had evolved to the point where missiles could strike the United States. This put the entire U.S. infrastructure at risk, including its information infrastructure.
The Survival Network and the Need for Packet-SwitchingPrior to Sputnik, the best way to protect vital information was to store it all in a centralized and heavily guarded location; the new possibility that that centralized location could be hit by a nuclear missile forced the Department of Defense to change its policy.
Instead of holding all the sensitive information in one location that could be targeted and destroyed, the U.S. began exploring ways to create a communications and storage network that could survive a strike.
A quick point has to be made here. It is a common misconception that the possibility of nuclear strike was the driving force behind the Internet. It wasn’t. As mentioned, the launch of Sputnik spurred the creation of ARPA, which is the department that created ARPANET. However, by the time they created ARPANET, their primary motivation was to share computer resources. That said, the threat of nuclear war was the primary reason that the RAND corporation - and specifically Paul Baran - came up with the theoretical concept of distributed networks.
In order to have an effective network that can survive a nuclear strike, data needs to be broken into small pieces that can be sent rapidly and reassembled with no errors where there is not a single point of failure. In short, you need packet switching. Like many conceptual inventions, its creation is shared between three people at three different organizations.
Paul Baran and Nuke-Proof NetworkIn 1959, Paul Baran was given the intellectual challenge of creating a communications system that would operate even if intermediate points were destroyed in a nuclear strike. The problem with the existing system was that it was centralized and worked in a linear fashion: A message left point A, traveled to the central switching area and was sent to endpoint B. Destroying point A or point B only stopped communication in one area, but destroying a central switching area took down the entire communications system.
The answer Baran came up with was to share communication among multiple, interconnected points. Each point in the communication system was connected to several other points in a non-linear arrangement. Baran referred to these points as nodes and used them to create distributed networks. He ran simulations of nuclear attacks on these distributed networks and found they could sustain major damage and still maintain end-to-end communications. Moreover, the more connections there were per node, the more resilient the network became as redundancies were created.
Message Blocks and PacketsThe key to getting information around this network was a routing doctrine known as “hot-potato” switching. Information would be divided into message blocks that were sent separately across the network. These blocks would be received by the nearest node and stored. That node would then look for the shortest path to send the message blocks to their destination node. If an area of the network was destroyed, the nodes would just send the message blocks around the missing areas.
The reason we don’t talk about message block switching is that Donald Watts Davies came up with the same concept at around the same time as Baran. Davies worked for the National Physical Laboratory (NPL) in the United Kingdom. Faced with a similar problem – how to create a robust communications network – Davies came up with packet switching. The team at ARPA eventually built such a network by drawing from both Baran’s and Davies’ work. They chose to call the system packet switching.