A. Cardenas, J. S. Baras, and K. Seamon
2006 IEEE Symposium on Security and Privacy, Oakland, California, May 21-24, 2006.
With the explosive growth of network techniques, in particular wireless communications, the traditional centralized, fixed networks can no longer satisfy the enormous demands on network connectivity, data storage and information exchanges. New types of networks, such as pervasive computing networks, mobile ad hoc networks and P2P networks, emerged in recent years in order to provide solutions for the increasing requirements on networked services. All those networks are autonomous networks, because they are distributed and self-organized. As a case study, we employ a specific application – distributed trust management – to understand and analyze the behavior and properties of these “anarchical” autonomous networks. We propose a statistical trust evaluation rule, prove its convergence and investigate its characteristics when the system is at the steady state. Our investigation results in several conclusions for the design of trust evaluation rules, some of which are unexpected if we do not have the stationary distribution at hand. Our study shows the importance and necessity of applying theoretical analyses to understand the complex characteristics of distributed, self-organized, autonomous networks.