An Approach for Fault Tolerance in Multi-Access Edge Computing
Liangjun Song, Gang Sun, Hongfang Yu · 2025
As demand for real-time applications and services continues to rise, multi-access edge computing (MEC) has emerged as a pivotal solution for achieving low-latency, highperformance computing. However, ensuring fault tolerance and service continuity in dynamic, complex network environments presents a significant challenge. This study introduces a service migration scheme for enhanced fault tolerance in MEC based on server classification, SMFT-SC, which classifies servers into critical and regular categories. The scheme allows for the proactive migration of critical services to healthy nodes before failures occur, thereby minimizing service disruptions and enhancing system reliability. SMFT-SC employs a multi-dimensional continuoustime Markov chain model to analyze the behavior of MEC components, providing insights that enhance both the reliability and performance of MEC systems. Furthermore, the scheme incorporates a hierarchical migration strategy that prioritizes the migration of critical services, reduces resource conflicts, and optimizes network efficiency. Simulation results demonstrate that the SMFT-SC scheme significantly reduces user access latency, improves service hit rates, and offers more robust fault tolerance across various scenarios, outperforming existing methods.