The optimization of replica distribution in the unstructured overlays

作     者：FENG GuoFu1,2 , LI WenZhong2, LU SangLu2, CHEN DaoXu2 & BUYYA Rajkumar3 1School of Information Science, Nanjing Audit University, Nanjing 210029, China 2State Key Laboratory of Novel Software Technology, Nanjing University, Nanjing 210093, China 3Department of Computer Science and Software Engineering, The University of Melbourne, Victoria 3010, Australia 

作者机构：1. School of Information Science Nanjing Audit University Nanjing 210029 China2. State Key Laboratory of Novel Software Technology Nanjing University Nanjing 210093 China3. Department of Computer Science and Software Engineering The University of Melbourne Victoria 3010 Australia 

出 版 物：《Science China(Information Sciences)》 (中国科学:信息科学(英文版))

年 卷 期：2012年第55卷第3期

页      面：714-722页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 081201[工学-计算机系统结构] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by National Basic Research Program of China (Grant No. 2009CB320705) National Natural Science Foundation of China (Grant Nos. 60803111, 61073028, 61021062) Jiangsu Natural Science Foundation (Grant Nos. BK2009396, BK2009100) 

主　　题：unstructured overlay item size distribution query rate success rate expected search size 

摘      要：Replication is a widely used technique in unstructured overlays to improve content availability or system performance. A fundamental question often addressed by previous work focused on: how many replicas ought to be allocated for each data item given the fixed query rates and limited storage capability? In this paper, we have put forth two optimal replica distributions to achieve the highest success rate and the lowest message consumption. Especially, we have investigated the influence of item size on replica distribution. Our results show that Square-Root Replication, which is traditionally considered to be optimal, is not always the best choice. Our study offers a new deep understanding of resource managment in self-organized unstructured overlays.