论文部分内容阅读
结合P2P和网格的特点,提出了从P2P与网格混合的分布式网络中选择超级节点的算法SSABC.算法使用网格信息监控系统(MDS)获取节点资源的动态属性信息,如可用带宽、空闲CPU、可用内存、当前连接数及在线时间等,根据以上属性计算节点的能力.当有新节点加入且超级节点均饱和时,从新节点或已加入节点中选择能力最高的作为新的超级节点.通过理论分析和仿真实验表明,基于能力选择的超级节点与随机选择的超级节点相比,提高了资源的查询成功率,缩短了平均查询跳数,并能够在超级节点饱和时均衡网络负载.当网络中节点数发生变化时,以上结论依然成立,说明了算法的可行性和稳定性.
Combined with the characteristics of P2P and grid, this paper proposes SSABC, an algorithm for selecting super nodes from a distributed network with P2P and grids.The algorithm uses Grid Information Monitoring System (MDS) to obtain dynamic attribute information of node resources such as available bandwidth, Free CPU, available memory, current connections and online time, etc. Based on the above attributes, the node’s capability is calculated.When a new node joins and both super nodes are saturated, the new super node is selected from the new node or the joined node with the highest capacity The theoretical analysis and simulation results show that compared with randomly selected super nodes, the super node based on capability selection improves the success rate of resource query, shortens the average query hops and balances the network load when the super nodes are saturated. When the number of nodes in the network changes, the above conclusion still holds, which shows the feasibility and stability of the algorithm.