Research

Northeastern's network security and distributed computing faculty wrestle with some of the most pressing questions in computer science today. Most of their projects focus on the burgeoning field of wireless and mobile computing, where optimizing resources is critical in ensuring that network applications run properly, particularly where security is concerned. Since resources such as battery power and bandwidth are limited over wireless networks, jamming due to a malicious attack or natural disaster could adversely affect health, safety, transportation, financial stability, and even lives.

So it's a true advantage having world-class experts on security (like Agnes Chan and Guevara Noubir) and distributed networks and algorithms (like Rajmohan Rajaraman and Ravi Sundaram) attacking these and other issues in search of innovative solutions. With a major grant from the Defense Advanced Research Projects Agency (DARPA), the network security and distributed computing faculty are working on a wireless anti-jamming project called Second-generation wireless Protocol Resiliency Enabled by Adaptive Diversification, or SPREAD.

Agnes Chan works on security protocols and authentication, looking at how resources are used to ensure secure communications. In a battlefield situation, troops continually send encrypted messages via airwave and update their shared wireless network keys to maintain security. Chan has devised efficient group key management schemes whereby the binary tree used to maintain the keys remains balanced at all times. Together with graduate student Jonathan Wong, she has designed an algorithm to update keys immediately, so that vulnerabilities resulting from lack of perfect synchronization can be reduced. She is also considering the synchronization problem of multicasting new keys, due to the dynamically changing memberships within groups of users. "International alliances are continually formed and broken, so you have to have the ability to instantly revoke or change keys," says Chan.

Guevara Noubir uses his expertise in the fields of security, physical-layer wireless communication, networking, and algorithms to help design the game-theory model for SPREAD. Specifically, he's working on analyzing the resiliency of existing communication systems, such as IEEE802.11 and GSM, to smart denial of service attacks. Noubir also focuses his research efforts on designing cross-layer protocols for robust and scalable heterogeneous wireless networks (for which he received an NSF CAREER Award). His ultimate goal is to develop the enabling technology for an ambient intelligence that is aware of people's presence, needs, and context, and is resilient to natural and man-made disasters, thus bridging the cyber and physical worlds.

Ravi Sundaram focuses on the algorithmic and game-theoretic aspects of secure communications. "We use a game-theory model and the question is: How do we strategize so we can always win?" explains Sundaram. "You need to know the strategies your opponent is going to use and the probability of when he or she is going to use them." His goal is to obtain practical methods and theoretical bounds for boosting the performance and security of communications in the presence of sophisticated adversaries who can jam transmissions. Sundaram has been able to show that highly robust communications can be achieved through strategies of adaptive diversification. He's working to devise efficient procedures to compute "the optimal strategies and associated probabilities for the communicating nodes."

Rajmohan Rajaraman approaches the problem from the perspectives of distributed computing and optimization under uncertainty. He's working to design protocols in which individual network nodes are able to cooperate and optimize for global objectives in adversarial environments, even with limited local information. "By formulating realistic mathematical models, we aim to develop protocols that work well not only under current known conditions, but also in future uncertain scenarios," says Rajaraman.

Such innovations are possible, in part, because members of the network security and distributed computing team have extensive experience in industry. Several have worked for North American and European telecommunications laboratories and others have participated in projects with the U.S. Department of Defense and MITRE Corporation. "We are experienced in coming up with theoretical foundations, and we also understand the problems in deployment," says Chan. "The knowledge of both theory and practice is our strong point."