Linear programming models for jamming attacks on network traffic flows

We present a new class of network attacks, referred to as flow-jamming attacks, in which an adversary with multiple jammers throughout the network jams packets to reduce traffic flow. We propose a linear programming framework for flow-jamming attacks, providing a foundation for the design of future protocols to mitigate flow-jamming. We propose metrics to evaluate the effect of a flow-jamming attack on network flow and the resource expenditure of the jamming adversary. We develop, evaluate, and compare a variety of flow-jamming attacks using the proposed metrics and the linear programming formulation. In addition, we formulate two approaches for distributed flow-jamming attacks for a set of jammers operating without centralized control and compare the performance to the centralized attacks using the linear programming formulation.