Research Area:  Edge Computing
Mobile Edge Computing (MEC) brings computation closer to user equipment (UE) and reduces the latency as well as energy consumption for computation offloading. However, with the development of Ultra-Dense Network (UDN), traditional offloading algorithms and mobility management (MM) approaches are not sufficient anymore with a frequent handover between base stations. It becomes far from optimal to execute offloading algorithms only at the beginning of a task when UE moves frequently or in the environment of UDN. Also, simple dynamic adjustment for offloading proportion of data input during UEs movement cannot solve this problem very well, because this approach executes global decision with information from only one position. Exploiting short-term mobility prediction in MM, we propose a novel dynamic mobility-aware partial offloading (DMPO) algorithm to figure out the amount of data for offloading dynamically, together with the decision of communication path in MM, minimizing the energy consumption while satisfying the delay constraint. This proposed algorithm predicts the time to next handover as well as the moving of UE in this phase and assigns data size to each time slot in this phase to achieve our goal of minimizing energy consumption while satisfying the delay constraint. In the simulation, we obverse the average delay and energy consumption with different delay constraint, moving speed of UE and base station density. The results demonstrate that our proposed algorithm saves up to 70% of energy while having a better performance in satisfying the delay constraint compared to traditional approaches.
Keywords:  
Author(s) Name:  Fangxiaoqi Yu,Haopeng Chen and Jinqing Xu
Journal name:  Future Generation Computer Systems
Conferrence name:  
Publisher name:  ELSEVIER
DOI:  10.1016/j.future.2018.07.032
Volume Information:  Volume 89, December 2018, Pages 722-735
Paper Link:   https://www.sciencedirect.com/science/article/abs/pii/S0167739X17329813