Aequitas: A 5G Scheduler for Minimizing Outdated Information in IoT Networks

Age of Information (AoI) is a promising metric to measure information freshness and optimizing AoI through scheduling is one of the most intensely studied areas in AoI research. To date, the vast majority of research on AoI scheduling has been based on simplified communication models that often fail to capture the complexities found in real-world network systems such as 5G. While there are some limited efforts on AoI scheduling that have ventured into exploring OFDMA-based data transmission models similar to those in 5G, they tend to neglect essential elements, such as channel-dependent Resource Block (RB) allocation and Modulation and Coding Scheme (MCS) assignment, rendering limited utility to real-world 5G systems. In this paper, we focus on developing 5G-compliant AoI schedulers. We study a specific problem with the objective of minimizing outdated information across all source nodes. This problem arises from practice where there is a specific information freshness requirement, known as AoI deadline for each source node. We present Aequitas, an innovative 5G scheduler designed to optimize this objective through joint RB allocation and MCS assignment, both of which are dependent on frequency and time-selective channel fading. We exploit a property called “uniform fairness," derived from the analysis of an optimal offline scheduler, to develop Aequitas. To meet stringent timing requirement in 5G, Aequitas leverages the parallel computing capability of a COTS GPU. Extensive evaluations demonstrate that Aequitas closely approaches the theoretical lower bound in terms of objective performance, while maintaining operational times below the 5G timing requirement.