|dc.description.abstract||We consider a wireless channel between a single pair of stations (sender and receiver) that is being “watched” and disrupted by a malicious, adversarial jammer. However, the jammer’s power is constrained by parameters $\rho$ and $\sigma$, that represent the rate at which the adversary can jam a packet, and the length of the largest bursts of jams, respectively. This, corresponds to the translation of the Adversarial Queueing Theory (AQT) constrains, to channel jamming.
The sender’s objective is to transmit as much useful data as possible, over the channel, despite the jams that are caused by the adversary. In this work, we focus on a simplified version of the problem, where given a transmission period $T$, the adversary is constrained in jamming up to $f$ packets. We believe it to be a building block for solving the general problem of AQT-based jamming.
We therefore develop deterministic scheduling algorithms that decide the lengths of the packets to be sent, in order to maximize the total payload successfully transmitted over period T in the presence of up to f packet jams, useful payload.
We first consider the case where all packets must be of the same length and compute the optimal packet length. Then, we consider adaptive algorithms; ones that change the packet length based on the feedback on jammed packets, and analyze them with respect to the simplified model, showing their optimality. Finally, we discuss how our solutions could be used to solve the more general problem.||