A father knows best: Vint Cerf re-thinks the Internet in Stanford talk

I've heard many of the "fathers of the Internet" say that security is the biggest flaw with the architecture of the modern IPv4 Internet, but not a lot of detail on why they think that is, or how it is they think that the network is going to mitigate the effect of billions of lines of code written without regard for security, much of it inherently exposed to attackers.

Ten years ago, one could reasonably argue that traceability was a key Internet security problem. A modern Internet would prevent people from launching anonymous attacks. When I was at Arbor networks, our most interesting competitor was a company called Asta, led in part by Stefan Savage, which had in their intellectual heritage a thesis on stochastic traceback of DoS attacks. Great stuff. But, as it turned out, not 3 years later the majority of non-trivial DDoS attacks were being carried out by organized armies of compromised machines. What's a secure Internet going to do about that problem?

I don't care about this as a practical matter. In reality, Vint Cerf's reimagining of the Internet has little force. The market (rightfully) determines what will be deployed. But it's a little sad to me if part of Cerf's logic is that the End to End Argument In Systems Design has outlived its usefulness --- that is, if he really believes we should concentrate more security intelligence into the network, instead of keeping the Internet dumb and insecure so people can innovate security from the edges.

I wasn't able to pull anything out of this article. Perhaps I read it too quickly. Did anyone get a sense of what sort of re-thinking he has in mind?

Something content-centric, perhaps?

http://en.wikipedia.org/wiki/Content-centric_networking

http://www.youtube.com/watch?v=8Z685OF-PS8 (an excellent google tech talk)

http://en.wikipedia.org/wiki/Van_Jacobson

Interesting. A link to the talk somewhere?