期刊
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
卷 67, 期 7, 页码 6533-6545出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TVT.2018.2808171
关键词
Mobile edge computing; virtual machine replicacopy; multiple applications; average response time; normalized total cost
资金
- National Natural Science Foundation of China [61771374, 61771373, 61601357]
- China 111 Project [B16037]
- Fundamental Research Fund for the Central Universities [JB171501, JB181506, JB181507, JB181508]
- Advanced and Applied Basic Research Projects of Chongqing [cstc2015jcyj40048]
Although mobile edge computing (MEC), as an extension of the cloud computing paradigm to edge networks, overcomes some obstacles of traditional mobile cloud computing, i.e., the reduced response time particularly, it is a nontrivial task to efficiently deploy virtual machine replica copies (VRCs) supporting multiple applications among numerous MEC servers in edge networks. To combat this issue, we are motivated to investigate in detail the optimal placement of VRCs to minimize the average response time (MART) in the MEC architecture with various requests demand among multiple applications and capacity constraints of MEC servers in edge networks. Besides optimal enumeration placement algorithm (OEPA) as benchmark, we design latency aware heuristic placement algorithm (LAHPA) with much lower computation complexity than that of OEPA. To enhance the performance of LAHPA on MART, clustering enhanced heuristic placement algorithm (CEHPA) is proposed, focusing on the optimal VRC placement in each cluster. We also develop substitution enhanced heuristic placement algorithm (SEHPA) to avoid falling into local optimal solutions. As corroborated by extensive simulation results, the performance of SEHPA on MART is very close to that of OEPA compared with LAHPA and CEHPA. Note that. CEHPA also outperforms LAHPA, and both are better than a general greedy placement algorithm. Furthermore, we evaluate the normalized total cost for services provision in edge networks, where SEHPA can also get more outstanding results than other algorithms.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据