RAP2P:一种基于资源广告的非结构化P2P系统

RAP2P: a resource advertising-based unstructured P2P system

研究了非结构化P2P系统的数据管理和相应的资源查询算法。提出了基于资源广告的非结构化P2P系统，即RAP2P。设计了一种类似谣言传播机制的资源信息广告算法AdGossip来传播共享资源广告，并且通过共享资源广告缓存限制机制来保证局部区域缓存空间对共享资源节点的最大覆盖率。分析和模拟结果表明，在获得100％查询命中率的条件下，RAP2P的资源定位消息开销约为泛洪查询的25％，查询的时延为泛洪查询的30％～50％，并且远远低于k-random walks。

The data management and resource locating algorithms of unstructured P2P system were studied. The proposed system was called RAP2P. The system simulated the marketing behavior, and shared the resource of advertised information to improve the resource locating performance. A gossiping-based resource advertising algorithm was proposed to propagate the resource information. By constraining the caching of resource information, the 2-hop search space stored the largest percent of all information. To exploit such resource layout, a dynamic shadow flooding search algorithm was presented. The analysis and simulation show that RAP2P requires about 25% message overhead of flooding algorithm and 30% to 50% query delay of flooding algorithm with 100% query hit rate, and its query delay is also far below that of k-random walks.