摘要
在对等计算应用中,副本复制技术是提升查询命中率、提升查询速度、维护负载均衡的一种有效方法,然而它也提升了在存储空间和流量上的代价。研究如何在结构化的P2P覆盖网中,通过拓扑优化手段减少复制中的冗余流量和冗余副本。首先在网络中选择支配集节点作为超级节点,设计一个层次化的、体现节点邻近度的P2P覆盖网,然后基于多hash函数,开发对应的复制技术以实现低代价的副本查询。该方法能够有效地在网络中分散副本,提升查询命中率,减少冗余消息和所需存储空间。给出了性能指标的理论分析,并通过仿真验证了该方法的优越性。
Replication technique is an effective method for improving the availability of data, enhancing performance in terms of query latency and load balance, while replication also brings significant costs in storage space and traffic. We studied how to decrease the redundant traffic cost for replication in structured P2P overlay by topology optimization. We developed a new Hierarchical Proximity-Aware P2P overlay with dominating set nodes acting as super peers, and designed a corresponding replication technique with multiple hash functions for costless query. Our method can efficiently disseminate replicas across the network, increase query hit ratio, and decrease redundant query messages and storage spaces required. We gave theoretical analysis for performance metrics,and by simulation,we verified the superiority of our method.
出处
《计算机科学》
CSCD
北大核心
2010年第7期87-90,共4页
Computer Science
基金
武汉市重点科技攻关项目(200750499172)资助