摘要
提出用分布式哈希表(DHT)为每台云服务器产生一个唯一的节点编号,该编号作为网络拓扑结构、检索信息存储和信息查询共同的标志符,从而形成一个适合分布式计算的结构化P2P覆盖网。设计了新的拓扑和路由协议来解决云资源的常数跳定位问题。仿真实验表明,经典的P2P算法平均查找跳数与网络规模成正相关,无法依据云计算的实际需要人为地控制查找跳数;该算法的平均查找跳数与网络规模无关,随着网络规模的增大而趋向于设定值,可以解决云资源的常数跳定位问题。
This paper proposed a method that using distributed hash table （DHT） to generate a unique node ID for each cloud server. The ID represented the common identifier for the network topology, the retrieval information storage and quel7. At last built a structure P2P overlay network for distributed computing. This paper designed a new topology and routing protocol to solve the cloud resources＇ constant jump locating problem. Simulation results show the classic P2P algorithms＇ average search hops is positively correlated to the network size, and it cannot control the average search hops according to the actual needs of cloud computing. However the proposed algorithm＇ s average search hops has nothing to do with the network size, and it tends to the set value as the network size increases, therefore it can ra^nlva tha r^nn^t~nt inmr~ ｜n^at;n Kl^m
出处
《计算机应用研究》
CSCD
北大核心
2013年第2期570-573,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(61170277)
上海市教委科研创新重点资助项目(12zz137)
关键词
云资源
云对等网络
拓扑
路由
资源定位
平均查找跳数
cloud resources
cloud peer-to-peer network
topology
routing
resources locating
average search hops
