Mean-Filed Theory and Macro Computation in Large-Scale Services Computing Systems with Resource Management and Job Scheduling
Services computing is an emerging and distributed computing mode,and it always becomes a hot and interesting research both from academe and from industry.Notice that the services computing systems display new characteristics,such as,stochasticity,loose coupling,concurrency,non-homogeneity,large scale,heterogeneity and so forth,thus the new characteristics is caused by some of technologies including SOA architecture,web services,cloud computing and internet of things.So far,this makes mathematical modeling and analysis of the large-scale services computing systems more difficult and challenging.From resource management and job scheduling in services computing,this paper establishes an integrated mechanism of realtime and dynamic control through combining the supermarket models,with the work stealing models,and with the scheduling of public resources.To this end,this paper proposes a new stochastic model for analyzing the large-scale services computing system.Because the state explosion is a basic difficulty,this paper applies the mean-field theory to proposing a macroscopic computational technique,which realizes to applying theory of nonlinear Markov processes to discussing the large-scale services computing systems and to providing their performance evaluation.Therefore,the methodology and results given in this paper can be applicable to dealing with more general large-scale services computing systems,including the resource management and the job scheduling.
Services computing Resource management Job scheduling Mean-filed theory Supermarket model Work stealing model Scheduling of public resources
Feifei Yang Yanping Jiang Quan-Lin Li
Northeastern University,Shenyang 110169 China School of Economics and Management Sciences Yanshan University,Qinhuangdao 066004 China
国际会议
河北秦皇岛
英文
118-121
2017-08-21（万方平台首次上网日期，不代表论文的发表时间）