TBTOA: A DAG-Based Task Offloading Scheme for Mobile Edge Computing

Abstract

Mobile Edge Computing (MEC) is an emerging computation paradigm that enables mobile devices to offload computation-intensive tasks to edge servers in order to speed up task processing. However, there are still a few challenging problems with MEC before it can be widely adopted. For instance, due to storage and computation constraints, only a limited set of services can be deployed on an edge server. If a mobile device requires a specific service, it can only offload its task to the edge servers that provide the corresponding service. In addition, a task might be composed of several dependent subtasks. The subtask that depends on other subtasks cannot be offloaded until the prerequisite subtasks are completed. Finally, due to the heterogeneity of edge servers, offloading a task to different edge servers could lead to varied energy consumption performance. Few studies consider both of these scenarios and to fulfill the gap, we investigate the task offloading problem in MEC under the dependency and service caching constraints. Specifically, we propose a heuristic algorithm named Table Based Task Offloading Algorithm (TBTOA), which is capable of predicting the impact of offloading decisions. Our experimental results indicate that TBTOA outperforms the existing offloading schemes for MEC in terms of makespan and energy consumption.