On the Age of Information in a CSMA Environment

In this paper, we investigate a network where $N$ links contend for the channel using the well-known carrier sense multiple access scheme. By leveraging the notion of stochastic hybrid systems, we find: 1) a closed-form expression of the average age when links generate packets at will 2) an upperbound of the average age when packets arrive stochastically to each link. This upperbound is shown to be generally tight, and to be equal to the average age in certain scenarios. Armed with these expressions, we formulate the problem of minimizing the average age by calibrating the back-off time of each link. Interestingly, we show that the minimum average age is achieved for the same back-off time in both the sampling and stochastic arrivals scenarios. Then, by analyzing its structure, we convert the formulated optimization problem to an equivalent convex problem that we find its optimal solution. Insights on the interaction between links and numerical implementations of the optimized Carrier Sense Multiple Access (CSMA) scheme in an IEEE 802.11 environment are presented. Next, to further improve the performance of the optimized CSMA scheme, we propose a modification to it by giving each link the freedom to transition to SLEEP mode. The proposed approach provides a way to reduce the burden on the channel when possible. This leads, as will be shown in the paper, to an improvement in the performance of the network. Simulations results are then laid out to highlight the performance gain offered by our approach in comparison to the optimized standard CSMA scheme.