一种拓扑感知分域聚簇的NICE协议改进算法  

Topology-aware Partition Clustering Improved Algorithm for NICE Protocol

作　　者：崔建群[1] 熊涛[1] 王博伦[1] 吴黎兵[2] 

机构地区：[1]华中师范大学计算机学院,武汉430079 [2]武汉大学计算机学院,武汉430072

出　　处：《小型微型计算机系统》2016年第11期2464-2468,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金面上项目(61170017;61272112;61370108)资助

摘　　要：NICE协议不能聚合相距较近节点,并且频繁分簇会进一步加剧这种情况,成为影响NICE协议性能的主要问题之一.提出一种具有拓扑感知的分域聚簇模型TPCM(Topology-Aware Partition Clustering Model),并根据此模型提出一种拓扑感知的分域聚簇的NICE协议改进算法.该模型能够将物理位置较近的节点聚簇在一起,成为域内节点,分簇时仅将域外节点进行分簇,从而实现数据包的就近传输;由于减少了参与分簇节点的总数量,因此也大大降低NICE协议的分簇次数,减少开销.试验结果表明,该模型可以大大降低NICE协议的分簇次数,有效降低组播树的传输时延,改善了协议的性能.Many neighbor nodes cannot be clustered in NICE protocol, and this situation is increased by frequent cluster split, it greatly reduces the performance of NICE protocol. This paper proposes a topology-aware partition clustering model （ called TPCM }, and furthermore,it proposed an Tl^M-based improved algorithm for NICE protocol. This Model can cluster the neighbor nodes, and just splits the foreign nodes, so it can localize the transmission of data. In addition It also can cut down the number of clusters split to reduce the expense due to the decrease the total of split nodes. The results of experiments show that TPCM can greatly reduce the number of split, obviously decrease the transmission delay, and improve the performance of NICE protocol.

关 键 词：应用层组播 NICE协议 拓扑感知 分域 聚簇 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]