版权所有:内蒙古大学图书馆 技术提供:维普资讯• 智图
内蒙古自治区呼和浩特市赛罕区大学西街235号 邮编: 010021
作者机构:Wuhan Univ Technol Dept Comp Sci Wuhan 430063 Hubei Peoples R China Natl Univ Def Technol Coll Comp Natl Key Lab Parallel & Distributed Proc Changsha Hunan Peoples R China Beijing Wuzi Univ Sch Informat Beijing 101149 Peoples R China
出 版 物:《FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE》 (下代计算机系统)
年 卷 期:2019年第100卷
页 面:921-937页
核心收录:
学科分类:08[工学] 0812[工学-计算机科学与技术(可授工学、理学学位)]
基 金:National Natural Science Foundation (NSF) [61672397, 61873341] Application Foundation Frontier Project of WuHan Beijing Intelligent Logistics System Collaborative Innovation Center Open Project Fundamental Research Funds for the Central Universities [2018-YS-063]
主 题:Edge cloud SaaS applications Replica migration
摘 要:The multi-replica strategy widely used in the traditional cloud makes storage costly and the replica cannot be accurately located. Due to many heterogeneous nodes in different locations of the edge cloud system, inefficient replica strategy will greatly affect the load balancing of the system. Therefore, the dynamic multi-objective optimized replica placement and migration strategies for SaaS applications in edge cloud are proposed in this paper. The multi-objective optimized replica placement problem is solved according to the fast non-dominated sorting genetic algorithm. In the experiment, the algorithm of replica placement will minimize the effective network utilization, reduce the response time and guarantee the load balancing of data nodes. Moreover, the replica migration model for access hotspots is proposed to obtain the pairing migration relationship from source node to target node. In the experiment, the replica migration algorithm will effectively reduce the migration time, minimize the response time and improve network resource utilization. Besides, the applicability of the proposed replica methods is shown in different scenarios such as the social media, intelligent agriculture, military field and transportation field. (C) 2019 Elsevier B.V. All rights reserved.