Was the Internet really designed to withstand nuclear war? In an earlier column, I characterized that description as an "urban legend."
A reader wrote in to disagree, noting that the Internet was funded by the Defense Advanced Research Projects Agency (DARPA) with the explicit goal of resistance to nuclear war. For validation, he urges folks to investigate the writings of some of the pre-eminent Internet developers of the time, particularly Paul Baran. (See this great short history of the Internet).
Who's right? Does it really matter? In my book, it matters a lot because understanding the process that leads to world-changing innovation is the only way to foster it. If we can't agree on how the Internet was created, how can we hope to do something like it again?
So did the government deliberately set out to create a nuclear-war-resistant network? Evidence suggests otherwise. It's certainly true that while at the RAND Corporation, Baran was actively engaged in the design of resilient communications systems intended to survive nuclear war. He was also one of the early proponents of a distributed architecture, and one of the co-inventors of packet technology - both of which are critical elements of the Internet's design.
But the actual architecture and creation of the 'Net was handled not by Baran, but by a team of researchers headed by Larry Roberts. Their goal was more modest: to effectively share computing resources among multiple organizations (including universities and government contractors). Roberts & Co. considered, but ultimately rejected, a centralized design for traffic management - not out of resilience concerns, but because nobody was willing to dedicate scarce and expensive computing resources to the problem of centrally managing and controlling traffic. Computer owners were, however, willing to spare a small fraction of their computational resources to route traffic in a distributed fashion - so Roberts opted to go with a Baran-like distributed design.
It's also true that original funding for the Internet came from DARPA and later still the National Science Foundation - all of which heavily funded Cold War initiatives. Back in the 1960s, most government-sponsored scientific funding was defense-related. Nothing new there - in the 1980s, I worked on a nuclear engineering project that was funded by President Reagan's Star Wars initiative. But the source of the funding implies nothing whatsoever about the goal of the project (my project had nothing to do with Star Wars, though we happily cashed the checks).
The bottom line is that while Cold War concerns definitely influenced the Internet's design, the shape of the 'Net is more a testament to economics than ideology. Most importantly, the design reflects not a top-down government-driven solution to a specific problem (resilient communications) but a bottom-up, government-supported solution to a more general problem: how to enable effective and economical sharing of computational resources.
The best part is another such burst of innovation is under way once more, with the same bottom-up, government-supported, pragmatic approach. This time it's called grid computing. Stay tuned.