Table of Contents Author Guidelines Submit a Manuscript
Mobile Information Systems
Volume 2016 (2016), Article ID 2784548, 17 pages
http://dx.doi.org/10.1155/2016/2784548
Research Article

Software Defined Resource Orchestration System for Multitask Application in Heterogeneous Mobile Cloud Computing

1State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China
2EBUPT Information Technology Co., Ltd., Beijing 100191, China

Received 20 January 2016; Revised 21 April 2016; Accepted 22 May 2016

Academic Editor: Alessandro Cilardo

Copyright © 2016 Qi Qi et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The mobile cloud computing (MCC) that combines mobile computing and cloud concept takes wireless access network as the transmission medium and uses mobile devices as the client. When offloading the complicated multitask application to the MCC environment, each task executes individually in terms of its own computation, storage, and bandwidth requirement. Due to user’s mobility, the provided resources contain different performance metrics that may affect the destination choice. Nevertheless, these heterogeneous MCC resources lack integrated management and can hardly cooperate with each other. Thus, how to choose the appropriate offload destination and orchestrate the resources for multitask is a challenge problem. This paper realizes a programming resource provision for heterogeneous energy-constrained computing environments, where a software defined controller is responsible for resource orchestration, offload, and migration. The resource orchestration is formulated as multiobjective optimal problem that contains the metrics of energy consumption, cost, and availability. Finally, a particle swarm algorithm is used to obtain the approximate optimal solutions. Simulation results show that the solutions for all of our studied cases almost can hit Pareto optimum and surpass the comparative algorithm in approximation, coverage, and execution time.