Mobility-Aware and Code-Oriented Partitioning Computation Offloading in Multi-Access Edge Computing

Multi-Access edge computing (MEC) enables less resourceful smart mobile devices (SMDs) to use computation and memory intensive applications by offloading them to edge servers with tolerant latency, significantly improving the computing paradigm of SMDs. But, when involving mobility in MEC, offloading strategy and overhead can be significantly influenced by the movements of SMDs. What(')s more, the movements of SMDs make it harder to deal with precedence among subtasks. However, to the best of our knowledge, few articles have studied mobility management in code-oriented partitioning offloading. To give an efficient solution, we propose the cost-saving offloading policy with mobility prediction using convex optimization and Lagrangian approach. Our scheme can help moving SMDs in MEC like driverless vehicles efficiently complete their tasks and reveal the impact of task dependency on completion time. The experimental results show that our algorithm can achieve at least 12% performance improvement on average than other three common methods.

Automated summary

Multi-Access edge computing (MEC) enables less resourceful smart mobile devices (SMDs) to use computation and memory intensive applications by offloading them to edge servers. However, mobility in MEC can significantly influence offloading strategy and overhead, making it difficult to handle precedence among subtasks. To address this, we propose a cost-saving offloading policy with mobility prediction using convex optimization and Lagrangian approach. Our scheme helps moving SMDs in MEC efficiently complete their tasks and analyzes the impact of task dependency on completion time. Experimental results demonstrate significant performance improvement compared to other methods.