Sharon Goldberg
Princeton University

3:00pm - 4:30 pm

Wu & Chen Auditorium

101 Levine Hall

" Securing Internet Routing"

 

Abstract

The Internet consists of multiple autonomous systems, each consisting of networks of devices that are prone to malfunction, misconfiguration, or attack by malicious parties, and each controlled by profit-seeking businesses with different economic goals. Despite these complex relationships, the Internet’s routing system currently operates under the assumption that all nodes in the network can trust each other. A decade of network security research has been dedicated to dealing with failures in this trust model; however, very few of these research proposals have been adopted in practice. In this talk I suggest that, before designing/deploying new network security protocols, we need to make a principled effort to understand the types of security guarantees that are possible within the engineering and economic constraints of the system. This talk describes several of our results in this vein.

We start by using game-theoretic (mechanism design) approaches to obtain a surprising negative result: Even the strongest known cryptographic Internet routing protocol ("Secure BGP") fails to provide certain important security guarantees in the presence of economically-motivated nodes. Motivated by these results, we use cryptographic approaches to tackle the problem from a different angle: We consider protocols that robustly detect and localize performance degradations in the presence of adversarial nodes. Our results challenge the conventional wisdom on a number of questions in network security, and suggest new directions related to the design of networks that can withstand selfish or adversarial behavior.

 

Bio:

Sharon Goldberg is a PhD candidate at Princeton University. Her research focuses on developing practical

solutions for problems in network security, by leveraging approaches from cryptography and game theory.

 

 

Tuesday, March 3, 2009
3:00 - 4:15
Wu & Chen
101 Levine Hall