title: Key Management for Long-Lived Sensor Networks

abstract:

Wireless networks consisting of low-power sensing devices are becoming a
ubiquitous part of the computing landscape. The emerging applications of these
sensor networks range from healthcare to warfare. Most of these applications
require protection of sensitive information. However, developing common
security services (confidentiality, authentication, etc.) for sensor networks
in a manner that meets the stringent resource constraints is a challenging
task. This challenge is further complicated by the ad hoc nature, intermittent
wireless connectivity, large scale, and unattended deployment of sensor
networks. As a result, sensor networks are highly vulnerable to physical node
capture and other attacks that may compromise critical security data, including
encryption keys used for confidentiality and authentication. Consequently, it
is necessary to provide sensor network key management services that, in
addition to being resource-efficient, are highly resilient to attacks.

We observe that sensor networks are inherently collaborative environments in
which sensor nodes self-organize and operate in groups that typically are
dynamic and mission-driven. Given this collaborative model, we propose an
efficient and scalable group key management solution to enable secure
communications in sensor networks. The solution is built upon Exclusion Basis
Systems; a combinatorial formulation of the group key management problem. In
this presentation, we describe a self-organizing key management scheme for
large-scale long-lived sensor networks consisting of low-end sensor nodes
clustered around more capable gateway nodes. The scheme distributes key
management functionality among multiple nodes. Using cluster-based
administrative keys, the scheme can localize the impact of attacks as well as
considerably improve the efficiency of maintaining freshness of communication
keys. Unlike most contemporary approaches, our scheme provides a mechanism to
overcome both multi-sensor and gateway captures. Moreover, the scheme enables
dynamic setup and management of arbitrary structures of secure group
communications, and supports both node eviction and addition. Analysis and
simulation results show that our scheme significantly boosts sensor network
resiliency to attacks while conserving network resources.


bio:

Mohamed Eltoweissy is an associate professor in the Bradley Department of
Electrical and Computer Engineering at Virginia Tech. He also serves as
Director of the Center for Cyber Assurance and Trust (CyCare). Eltoweissy
received his PhD from Old Dominion University and his MS and BS from Alexandria
University, Egypt in 1993, 1989, and 1986 respectively. Eltoweissy has over 80
publications in archival journals and respected books and conference
proceedings. Among Eltoweissy^Òs research contributions are  novel
combinatorial-based survivable key management schemes for sensor and ad hoc
networks, service-centric architecture for sensor networks, stochastic models
for the optimization of security protocols, and denial-of-service resilient
routing architecture for mobile ad hoc networks. Eltoweissy is also active in
serving on program committees and NSF panels, in journal editorials and
organization of professional meetings. Eltoweissy is a senior member of IEEE,
and a member of ACM, ACM SIGBED, and ACM SIGSAC. In 2003, Eltoweissy was
nominated for the Virginia SCHEV outstanding faculty awards; the highest honor
for faculty in Virginia.