Resource management for heterogeneous services in cognitive mesh networks

In this paper, we propose a novel end-to-end resource management scheme for heterogeneous services in cognitive mesh networks (CMNs). Due to the highly dynamic nature of spectrum availability in cognitive radio networks (CRNs), we design a two-step approach, namely long-time route selection and periodic resource allocation. By introducing the interference cushion to resource allocation, we prevent real-time (RT) traffic from starving non-real-time (NRT) traffic. We formulate the problems of channel and power allocation for RT and NRT flows as two mixed-integer nonlinear programming (MINLP) problems which are NP-hard in general, and propose low-complexity yet efficient suboptimal algorithms to solve the formulated MINLP problems. Simulation results show the effectiveness of the proposed algorithms in terms of RT flow´s mean access ratio and NRT flow´s mean throughput. Also, we find that how to set interference cushion can affect the performance tradeoff between QoS provisioning for RT traffic and throughput maximization for NRT traffic.