SINR-based MAC Games for Selfish and Malicious Users

We study the selfish and malicious transmitter behavior in medium access control (MAC) layer of wireless networks with a general channel model based on the signalto-interference-plus-noise-ratio (SINR) achievable by multiple transmitters at a single receiver. We formulate the denial of service attacks in MAC operation as a non-cooperative game of selecting the individual transmission power levels and evaluate the effects of malicious transmitters that have the objectives of jamming the transmissions of selfish nodes. We consider performance objectives of throughput and power efficiency as well as impose power and SINR constraints that result from the hardware restrictions and delay efficiency objectives. Our primary focus is the case of one selfish and one malicious transmitter, whereas the analysis is also extended to arbitrary number of transmitters. We show the existence and uniqueness of the non-cooperative Nash equilibrium strategies for two different classes of utility functions that represent the linear combination or ratio of successful transmission rate and power consumption. We also discuss the social equilibrium strategies that maximize the total system utility. For distributed operation, we introduce adaptive best-response update mechanisms that converge to the non-cooperative equilibrium. Finally, we consider an alternative successful packet reception criterion based on SINR threshold and show that the existence and uniqueness of Nash equilibrium strongly depend on costs, rewards and constraints.