Primsp: Mobility-Aware Service Placement Strategy Based on Service Priority

Mobile edge computing enhances cloud computing by providing low-latency and real-time response services at the network edge. However, the limited and heterogeneous resources of edge servers pose challenges for efficient service placement in multi-user scenarios. To address this issue, we propose a mobilityaware service placement strategy based on service priority (PriMSP), which leverages a priority mechanism based on service sensitivity to mitigate migration conflicts and minimize service response time. Specifically, we formulate an optimization model aimed at minimizing the service response time while considering user mobility and constraining service migration costs. To reduce frequent service migrations and conflicts caused by user mobility, we introduce multi-step trajectory prediction and a prioritybased processing strategy to optimize service migration and mitigate migration conflicts, respectively. Furthermore, by integrating service priority and trajectory prediction, we propose a proactive service placement method based on deep reinforcement learning (DRL). We conduct experiments in a simulated environment using a real dataset, and the results demonstrate that our method effectively reduces the service latency of latency-sensitive applications while considering migration costs.