Minimizing Age of Information for Multiple Flows

Abstract

Age of Information (AoI) is an emerging metric measuring the freshness of a flow at its destination, which is critical for time-sensitive applications. We consider multiple flows served by a single server. To minimize the average AoI over all flows, we formulate preemptive and non-preemptive Maximum Age Difference (MAD) algorithms, and experimentally observe their superior performance with respect to various other scheduling algorithms for multi-flow networks. We also observe that both MAD and MAF (Maximum Age First) respond favorably to flow diversity, with a reduction in AoI.