论文部分内容阅读
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.
NICE protocol can not converge nodes with close neighbors, and frequent clustering will further aggravate this situation and become one of the main problems that affect the performance of NICE protocol.A topological-Aware Partition Clustering (TPCM) Based on this model, an improved algorithm of NICE protocol for topologically aware sub-cluster is proposed. This model can cluster the nodes closer in physical location into nodes in the domain, Which reduces the number of clusters in NICE protocol and reduces the overhead greatly.Experimental results show that this model can greatly reduce the clustering of NICE protocol Reduce the transmission delay of the multicast tree, and improve the performance of the protocol.