Throughput analysis of channel surfing in jammed single-hop wireless networks

We develop a theoretical approach to derive system throughput for single-hop wireless networks where U users share a spectrum set θU of n independent channels, and J jammers attack the system by selecting channels from a spectrum set θJ of N independent channels while θU ⊆θJ. Users and jammers follow the given policies to randomly select channels at each time slot. The whole system can be modeled as a Markov chain. The key point is to derive the transition probability matrix and obtain the stationary distribution. Then, the theoretical system throughput can be computed straightforwardly. Another interesting problem is to study the system failure probability for the scenario where each user pair has multiple links. The network model is modified as: m user pairs share a spectrum set θU and each user pair has c links. System failure happens when at least one user pair fails all of its links. A two-step approach is developed to compute the system failure probability for the modified model. Numerical experiments are used to validate theoretical results.