| dc.description.abstract | In this work we consider the communication over a wireless link, between a sender and a receiver, being disrupted by a jammer. The objective of the sender is to transmit as much data as possible to the receiver in the most efficient way. The data is sent as the payload of packets, and becomes useless if the packet is jammed. We consider a jammer with constrained power, defined by parameters \rho and \sigma, which represent the rate at which the adversary may jam the channel, and the length of the largest burst of jams it can cause, respectively. This definition translates to the Adversarial Queuing Theory (AQT) constraints, typically used for packet arrivals.
We propose deterministic algorithms that decide the length of the packets sent in order to maximize the goodput rate; i.e., the amount of useful payload successfully transmitted over time. To do so, we first define and study a static version of the problem, which is used as a building block for the dynamic problem. We start by assuming packets of the same length and characterizing the corresponding quasi-optimal length. Then, we show that by adapting the length of the packets, the goodput rate can be improved. Hence, we develop optimal adaptive algorithms that choose the packet lengths depending on the jams that have occurred up to that point in time, in order to maximize the total
payload transmitted successfully over a period T in the presence of up to f jams. | |
