A fast hybrid multi-site computation offloading for mobile cloud computing. (15th February 2017)

• Record Type: Journal Article
• Authors: Goudarzi, Mohammad; Zamani, Mehran; Haghighat, Abolfazl Toroghi
• Is Part Of: Journal of network and computer applications. Volume 80(2017)
• Page Start: 219
• Page End: 231
• Publication Date: 2017-02-15
• Subjects: Mobile cloud computing -- Multi-site computation offloading -- Optimal partitioning -- Near-optimal partitioning -- Optimization; Microcomputers -- Periodicals; Computer networks -- Periodicals; Application software -- Periodicals; Micro-ordinateurs -- Périodiques; Réseaux d'ordinateurs -- Périodiques; Logiciels d'application -- Périodiques; Application software; Computer networks; Microcomputers; Periodicals; 004.05; 004
• DOI: 10.1016/j.jnca.2016.12.031
• Languages: English
• ISSNs: 1084-8045

Abstract

Abstract: Due to nowadays advances of mobile technologies in both hardware and software, mobile devices have become an inseparable part of human life. Along with this progress, mobile devices are expected to perform various types of applications. However, the energy challenge of mobile devices along with their limited computation power act as barriers in front of this expectation. To address this deficiency, mobile cloud computing has been proposed in which computation-intensive tasks of mobile devices are offloaded to cloud resources, so that the energy and time spent for executing the mobile application are reduced. Moreover, in the real world, different types of clouds/servers with heterogeneous processing speeds and access delays are available for offloading resulting in time consuming process of making decision for offloading. To address these objectives, we propose a fast hybrid multi-site computation offloading solution which finds the offloading solution in a timely manner. Our solution includes two decision algorithms to achieve the optimal or near-optimal solution of multi-site offloading considering the application size. We evaluated the efficiency of the proposed solution using both simulation and testbed experiments. The evaluation study demonstrates that our proposal can outperform existing optimal and near-optimal counterparts in terms of weighted execution cost, energy consumption and execution time.