A Theoretical Framework for Random Access: Effects of Carrier Sensing on Stability

At the core of stability analysis of random-access networks is the characterization of the stability region of input rates and the operating region of transmission probabilities for achieving stability. In Dai (2022), a theoretical framework was proposed to tackle the above long-standing issues for the sensing-free Aloha case. In this paper, the analysis is extended to Carrier Sense Multiple Access (CSMA) to demonstrate the effects of carrier sensing on the stability performance. CSMA fundamentally differs from Aloha in that nodes need to sense the channel, which takes time. Meanwhile, they are closely related as given that the channel is idle, the contention process of nodes in a CSMA network is identical to that in an Aloha network. That leads to an interesting mapping between the stability regions with CSMA and Aloha. The analysis shows that with CSMA, the service rates of nodes’ queues, stability region of input rates, and complete operating regions of initial transmission probabilities for given backoff function are all crucially determined by the normalized sensing time and the failed transmission time. When the normalized sensing time is large, the stability region could be smaller than that with Aloha, indicating that carrier sensing may not always be beneficial for short-packet transmissions.