The paper the article references can be found here: http://netlab.caltech.edu/pub/papers/Doyle-topology-PNAS-0508.pdf I've several good discussions on this topic with David Alderson, one of the co-authors, and they understand this area as well as anyone. The paper and their previous ones definately debunk that the Internet is not successfully descibed by the scale free model, but I still have not seen a refutation of the empirical work. Still need to read the latest paper, but their previous work was based on simulations of network structures, or samples of the Abeliene network. While the SF model fails, I have not seen an argument to contracdict that when hub routers were failed on empirical (real world) data the network did not degrade gracefully. It could be that the data used to create the router network was flawed by various sampling techniques, a point David has made in the past, but have not seen an empirical study with "better" or "accurate" data from the whole Internet that refutes the earlier studies. I'd guess that the Cal Tech team makes the case the prior work was so flawed there is nothing to refute, and we are starting from zero and their approximations are a much cloesr starting point. Still it would be interesting to have a real world accurate global router data set to test against. This probably gets back to information sharing issues, but intersting paper. Thanks for the link. sean ----- Original Message ----- From: Sean Donelan <sean@donelan.com> Date: Tuesday, October 4, 2005 11:26 am Subject: The Internet is not scale-free
The Internet may not be as vulnerable to centralized attacks as previouslysuggested, researchers in California reported Monday.