It is frequently stated that the ARPANET was developed in order to provide an iron-clad, bullet-proof, nuclear-holocaust-resistant computer communications network for the US Armed Forces in the event of a Soviet attack.
According to the book Where Wizards Stay up Late..., this is not true. Appropriately for open-source software, the ARPANET was developed to scratch an itch: J. C. R. Licklider, formerly of Bolt, Beranek and Newman and now the head of Behavioral Sciences and Command and Control programs at ARPA, found it frustrating to have three different terminals on his desk, connected to three different computer systems, with three different login procedures. The ARPANET was developed in part, then, as a way of getting around the inconvenience of having to maintain separate access protocols to the computers of the time, which, as I explained in my earlier post, were tremendously (one might even say "ridiculously") varied and incompatible.
The possibility of making the Internet "self-healing" by increasing the number of point-to-point links, and thereby creating a network where there would always be a working path between any two sites, was discussed, but ultimately rejected as being too demanding on resources:
The question remained: How much redundancy in the interconnections between neighboring nodes would be needed for survivability? "Redundancy level" was [RAND researcher Paul] Baran's term for the degree of interconnectivity between nodes in the network. A distributed network with the absolute minimum number of links necessary to connect each node weas said to have a redundancy level of 1, and was considered extremely vulnerable. Baran ran numeous simulations to determine the probability of distributed network survival under a variety of attack scenarios. He concluded that a redundancy level as low as 3 or 4 - each node connecting to three or four other nodes - would provide an exceptionally high level of ruggedness and reliability.... Even after a nuclear attack it should be possible to find and use some pathway through the remaining network.
As the IMP Guys [a nickname for the people at BBN responsible for writing the software that would control the Interface Message Processors, the ARPANET's equivalent of Internet routers] laid out their plans in Washington, it became appparent that the ARPA network would be a hybrid of the original ideas of Baran and Davies. The ARPA network would use an adaptive routing scheme that the IMP Guys had developed on their own, but which was similar to the basic idea that Baran had sketched. However, unlike Baran's theoretical network, the ARPA network would not have nearly the same redundancy level or number of links to each node. Nodes in the BBN scheme were normally linked to two neighboring nodes occasionally to one or three. As it was now conceived, just two failed links could divide, or partition, the network into isolated segments. Paul Baran had said that a network with a multitude of redundant links could be built of inexpensive parts; if a part malfunctioned, there would be others to back it up. The low level of redundancy in the ARPA network was closer to Davies' plan. [Frank] Heart's approach was to make every part of the network as robust and reliable as possible.
Overall, I think the concept of the Internet works quite well even without the extra reliability built into the Baran plan - most of the time; it's also a lot cheaper, too. I suspect that the Internet would have been much more difficult to sell to consumers - if not everyone else, also - if every dialup or ADSL line required not one, but four phone lines!
Where Wizards Stay up Late: The Origins of the Internet, Katie Hafner and Matthew Lyon. Touchstone, 1998. p59
ibid, pg 113.
NB: The previous article on this subject incorrectly attributed the authorship of the book Where Wizards Stay up Late... to Larry Roberts, Frank Heart and Bob Kahn. This was incorrect; the aforementioned persons were not the authors of the book, but were mentioned in the book as being instrumental in the development of the ARPANET. Apologies to the real authors of the book, Katie Hafner and Matthew Lyon, who have been correctly attributed in references to the book included in this article.
Excluding quoted portions copyrighted by third parties, all text in this post is copyright (c) 2006 Jeff Rollin. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.