Mobile Computation Offloading Strategy Based on Static Information and Dynamic Partition

This paper presents a mobile computation offloading strategy, a novel framework which combines the static information and the dynamic partition to achieve low latency and energy cost. Previous works can lead to either high resource cost or inaccurate offloading decisions. In the proposed method, the static information is introduced into the strategy establishment process. In static information extraction, two offline strategies are established with the best and worst predicted communication quality. Then, strategies are compared with each other to find the same decisions, and every component is labeled as non-removable, removal or removable; in dynamic partition, the removable components are allocated into mobile terminal and cloud with the practical communication condition. Additionally, a linear time search method is proposed to find the optimal partition of application. To evaluate the strategy performance, three applications are used to test the efficiency of the strategy. The experiment demonstrates that the proposed strategy enables more resource saving in energy cost and latency than existing methods.