Analyzing how to deal with new-node that does not match the super-node in super-node P2P network based on leader-follower algorithm can help improve the efficiency and performance of super-node. The paper introduced general class node and splitting algorithm, and the nodes that do not match every super-node were managed by the general class node. When the nodes reached a certain number, the splitting algorithm was used to split these nodes into several semantic similarity clusters. Finally, the merge sorting algorithm chose the optimal node as super-node. The experimental results show that the proposed method improves the efficiency and performance of super-node, and it has good feasibility.%基于leader-follower算法的超级节点P2P网中,研究如何处理新进节点与各超级节点语义不匹配问题,有利于提高节点匹配效率和超级节点性能.引入通用类节点和分裂算法,将与各超级节点语义不匹配的新节点交由通用类节点管理,当管理的节点数目达到一定规模后,采用分裂算法将其分裂为若干语义相似簇,最后用合并排序算法从中选择最优节点作为超级节点.实验表明所提方法提高了节点匹配效率和超级节点性能,具有良好的可行性.
展开▼