Robust Functionality and Active Data Management for Cooperative Networks in the Presence of WMD Stressors
Majeed M. Hayat, Patrick G. Bridges, Yasamin Mostofi, Patricia Crowley · 2011
Abstract : In this project , we have begun the development of a rigorous probabilistic framework enabling new understanding and control of distributed networks vulnerabilities to WMD-induced failures. We have developed a general stochastic queuing model and performed basic analysis to characterize the statistics of the task completion time treated as a random variable. Using the developed framework, we have established a theoretical/computational optimization tool that maximizes a network's robustness to node/link failures by ponstantly redistributing the network' s computational loads. The framework also includes a rigorous treatment of the time evolution of networks subject to multiple random-in- space-and-time disruption and restoration events. Next, we have developed a theory to describe and solve the problem of a network attempting to reach consensus on the occurrence of a WMD attack based upon network data . To this end, we have developed the mathematical foundations of networked binary consensus over noisy links. Finally, we have considered real-time adaptation in distributed, unreliable networks irl the face of the potentially correlated failures induced by WMD tressors, establishing methods for driving adaptation using low-overhead monitoring systems.