QoS-Constrained Service Migrations in MEC Using Dynamic Service Prioritisation Algorithm

The rapid expansion of 5G networks and the increased focus on user experience has resulted in the growth of Multi-access Edge Computing (MEC) at the edge of wireless networks. The indeterministic workloads handled by MEC often lead to overloaded applications. Services migrations are performed in MEC to manage the application performance, potentially minimising the impact on the user's Quality of Experience (QoE). While prior research on application migrations in MEC has primarily focused on subscriber mobility, energy management, and resource optimisation, a gap exists in understanding how these migrations impact application performance in MEC environments with stringent Service Level Agreements (SLAs). This research proposes a new approach for service migrations in MEC based on dynamic service prioritisation. The proposed Dynamic Service Prioritisation Algorithm (DSPA). uses a variable application performance watermark to trigger migrations, enabling MEC infrastructure to dynamically adapt to indeterministic workloads. DSPA selects services for migration by optimally balancing service load and migration time parameters. Through extensive simulations, our study proves that DSPA is best suited for managing the performance of diverse applications hosted in MEC. DSPA exhibited superior performance in minimising SLA violations and mitigating user experience impact during service migrations compared to other contemporary algorithms.