Nodes Clustering Method in Large-Scale Network

Nodes in large-scale network are always divided into many clusters in order to facilitate network management. Super-node based P2P network is an example. The physical scope and the number of clusters are two conflicting objectives and clustering algorithm should strike a balance between them. In this paper, nodes clustering problem is formally described and pre-optimized k-means algorithm is put forward. First pre-optimization algorithm is used to determine the number of clusters and determine reasonable initial cluster centers, then k-means algorithm is used to achieve clustering result. The algorithm presented has low complexity and simulation results shows its effectiveness.