双层云环境下面向移动应用的最优服务分配算法  

Optimal service allocation algorithm for mobile applications in 2-tied cloud environment

作　　者：黄会群[1,2] 

机构地区：[1]中南大学公共卫生学院,长沙410078 [2]湖南财政经济学院信息管理系,长沙410205

出　　处：《计算机应用研究》2015年第12期3750-3755,共6页Application Research of Computers

基　　金：湖南省教育厅科研项目(12C0561)

摘　　要：对双层云环境下的移动应用进行研究,提出一种新的框架,将移动应用建模为任务的位置—时间工作流(LTW)。LTW流对分层移动云资源的最优映射是个NP难题,进而提出一种称为Mu SIC的高效启发式算法。基于真实的云进行仿真实验,结果表明,Mu SIC支持运行的可扩展性,同时提高了Qo S水平。相比于只使用公共云的算法,延时和功耗降低了25%,成本下降了35%。另外,在不同的模式下Mu SIC均具有优越的性能,在预测移动用户地点—时间工作流时面对差错/不确定性也具有较好的鲁棒性。This paper proposed a novel framework by the research on the mobile application in 2-tied cloud. It modeled mobile applications as the location-time workflows （ LTW ） of tasks. The optimal mapping of LTWs to tiered mobile cloud resources was an NP-hard problem. It proposed an efficient heuristic algorithm called MUSIC. The experiments based on real world clouds were done. The results indicate that MUSIC supports scalable operation while improving QoS. It observed about 25% lower delays and power about 35% decrease in price in comparison to only using the public cloud. In addition, MUSIC performs quite well under different mobility patterns, and is resilient to errors/uncertainty in prediction of mobile user locationtime workflows.

关 键 词：分层云 移动应用 位置—时间工作流 最优映射 NP难题 功耗 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]