摘要
为了有效实现结构化P2P系统中数据均衡分布,借鉴并行数据库中数据划分的基本思想,通过在节点加入和数据加入时的存储均衡算法实现大数据量在系统中存储均衡,使得系统存储差异系数大大降低。设计了一个使用数据划分的结构化P2P模型Balance-Peer。在不需要全局信息的情况下,实现动态数据划分方法。实验结果表明了该存储均衡策略是有效的。
In order to realize the balanced distribution of data in distributed systems, the concept of data partition in parallel data- bases is used to realize the load balance in P2P systems. When the node and the data join, the storage balanced algorithm is exe- cuted to achieve the storage balance. The whole system greatly reduces the storage variation coefficient. A structured P2P model called balance-peer is designed, and a method is introduced to partition the data rang without any global information on the load. The result show that our storage balance strategy is valid.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第4期1153-1156,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60703007)
首都经贸大学科研基金项目(2012XJG021)
北京自然科学基金项目(9123025
9122003)
