A Unified Resource Allocation Framework For Defending Against Pollution Attacks In Wireless Network Coding Systems

Wireless network coding systems can significantly improve the bandwidth effi-ciency in wireless networks. However, a type of attack, called pollution attack, can cause severe damages in network coding systems and thus make the deployment of network coding systems impractical. Many approaches have been proposed to defend against pollution attacks. However, the current approaches implicitly assume that the defender has adequate resources to defend against pollution attacks. In practice, the resources of the defender are often limited, and the previous works provide no infor-mation for the defender to allocate the resources to get better defense performance.This thesis considers the case where the defender’s resources are limited and pro-poses a resource allocation framework for the defender to defend against pollution at-tacks. This thesis first studies this problem and proposes approaches for one-session transmissions. A two-player strategic game to model the interactions between the de-fender and the attacker. Under this model, corresponding algorithms are proposed to find the best strategy for the defender. Then, this thesis extends the study of the resource allocation problem to a multi-session setting. This thesis proposes an extensive game model and an enhancement algorithm to solve resource allocation problem under this circumstance. Furthermore, as a part of our enhancement algorithm, a novel malicious node identification mechanism is proposed in this thesis. Finally, thesis conducted ex-tensive simulations to evaluate the proposed framework. The results demonstrate that our algorithms can significantly improve the utility of the defender, with reasonable computation time.