VQ-CSMA: Throughput-Optimal Low-Delay Random Access

Abstract

Queue based carrier-sense multiple access (Q-CSMA) is designed to ensure throughput optimality in wireless networks, where the access probability depends on the local queue length information. However, some links may be long-term blocked if their neighbors continuously occupy the channel, which is called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">starvation</i> and leads to poor delay performance. In this letter, to deal with the starvation problem, we construct a neighbor-aware virtual queue system, which considers not only the local queue lengths but also the neighbors’ queue lengths. In the virtual queue system, we design a new Lyapunov function and derive a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">virtual queue based CSMA (VQ-CSMA)</i> algorithm. With the presence of negative virtual queues, we discuss four cases of virtual queues and prove that VQ-CSMA achieves throughput optimality by analyzing the corresponding maximum-weight scheduling. The simulation results demonstrate that VQ-CSMA alleviates the starvation problem and consequently leads to a better delay performance compared to existing Q-CSMA.