A Multi-Objective Compounded Local Mobile Cloud Architecture Using Priority Queues to Process Multiple Jobs

• Published in: PloS one Vol. 11; no. 7; p. e0158491
• Main Authors: Wei, Xiaohui, Sun, Bingyi, Cui, Jiaxu, Xu, Gaochao
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 15.07.2016; Public Library of Science (PLoS)
• Subjects: Algorithms; Analysis; Architecture; Biology and Life Sciences; Cloud computing; Cloud Computing - economics; Collaboration; Computer and Information Sciences; Computer architecture; Computer science; Computer simulation; Computers, Handheld; Data centers; Electronic devices; Engineering and Technology; Information Storage and Retrieval - economics; International conferences; Internet; Iterative methods; Mobile computing; Mobile devices; Multiple objective analysis; Physical Sciences; Portable equipment; Queues; Research and Analysis Methods; Researchers; Scheduling; Smartphones; Social Sciences; Software - economics; Software Design; Task scheduling; Time Factors; Wireless networks; China
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0158491

As a result of the greatly increased use of mobile devices, the disadvantages of portable devices have gradually begun to emerge. To solve these problems, the use of mobile cloud computing assisted by cloud data centers has been proposed. However, cloud data centers are always very far from the mobile requesters. In this paper, we propose an improved multi-objective local mobile cloud model: Compounded Local Mobile Cloud Architecture with Dynamic Priority Queues (LMCpri). This new architecture could briefly store jobs that arrive simultaneously at the cloudlet in different priority positions according to the result of auction processing, and then execute partitioning tasks on capable helpers. In the Scheduling Module, NSGA-II is employed as the scheduling algorithm to shorten processing time and decrease requester cost relative to PSO and sequential scheduling. The simulation results show that the number of iteration times that is defined to 30 is the best choice of the system. In addition, comparing with LMCque, LMCpri is able to effectively accommodate a requester who would like his job to be executed in advance and shorten execution time. Finally, we make a comparing experiment between LMCpri and cloud assisting architecture, and the results reveal that LMCpri presents a better performance advantage than cloud assisting architecture.