An Online Robust Resource Allocation Algorithm for Mobile Edge Computing

This letter studies fault tolerance and resource allocation for offloaded tasks in Mobile Edge Computing (MEC). The limitations of computational resources and wireless transmission power, coupled with server failures in MEC, present challenges for optimizing resource allocation and ensuring uninterrupted task execution. Existing literature primarily addresses resource allocation and fault tolerance separately, resulting in resource wastage and an increased count of faulty tasks. To address these issues, we first propose a novel method for jointly allocating computational resources and wireless transmission power for offload task. We then introduce the Dynamic Fault Tolerance Task Packing (DFTTP) algorithm, an online task packing algorithm. DFTTP, based on a new variant of the primary-standby replica mechanism, can tolerate multiple server failures simultaneously. Theoretical analysis and simulation results demonstrate that DFTTP, together with the proposed resource allocation method, can reduce the number of faulty tasks and enable higher resource utilization compared with existing algorithms.