摘要
针对无结构P2P网络中稀有资源搜索成功率低、搜索代价高的问题,提出2种基于网络覆盖的稀有资源搜索策略:RSR和FRSR。RSR在随机漫步的基础上通过考虑邻居节点的热度改进请求转发方式,FRSR通过结合洪泛搜索改进随机漫步转发策略。实验结果表明,RSR搜索稀有资源的时间比普通随机漫步减少了22.9%,平均搜索成功率提高了26.2%,通信开销降低了22.8%,FRSR比随机转发方式减少了15.4%的搜索时间,提高了14.2%的搜索成功率。
Aiming at the problem that locating rare resource is time-consuming with low success rate in unstructured P2P networks, this paper proposes two strategies for rare resource searching: Random Search for Rare resource(RSR) and Flooding and Random Search for Rare resource(FRSR). Based on random walk, RSR modifies request-forward manner by taking the popularity of neighbor nodes into consideration on passing searching request. FRSR combines flooding search with random walk. Simulations show that compared with Random Walk, RSR can reduce response time and communication cost by 22.9% and 22.8%, and improves success rate by 26.2%. Compared with random request passing method, FRSR reduces the response time by 15.4% and improves success rate by 14.2%.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第11期68-70,共3页
Computer Engineering
基金
国家自然科学基金资助重点项目(60533020)
教育部博士点基金资助项目(20060358011)