Age-of-Information Minimization in Multi-Server Edge Computing Networks for URLLC Services

The evolution towards 6G communication technology heightens the demand for data freshness. Consequently, Age of Information (AoI), a key metric quantifying data freshness, has raised significant attention from academia and industry. This paper investigates a Multi-access Edge Computing (MEC) network with multiple servers designed to support mission-critical, low-latency computational services. To meet stringent latency requirements, we employ finite blocklength (FBL) codes for data transmissions during the task offloading. Our focus is on minimizing the average AoI of the MEC system via system scheduling, which is particularly crucial for recently emerged applications such as health monitoring systems, automatic driving systems and intelligent factory inspection system. We characterize the transmission reliability with FBL codes in the communication phase. Using extreme value theory, we analyze the occurrence of extreme queue length violations during the computation time phase. Based on the characterizations, we develop an optimal framework incorporating server selection and scheduling strategies for minimizing the average AoI. Via numerical simulations, we validate our algorithm's effectiveness in enhancing AoI performance, demonstrate how varying parameters affect system performance, and illustrate the potential of our method for guiding future MEC system designs.