Toward efficient distributed network monitoring

Efficient and scalable management of large networks requires distributed management systems. In this paper, we present two distributed algorithms for network monitoring. Monitoring forms the basis of control and management systems. Effective monitoring ensures the network operates within desirable parameters. Conducting such monitoring introduces communication overhead that may reduce the overall effective throughput. The monitoring consists of measuring properties of the network, and of inferring an aggregate function from these measurements. We develop two distributed monitoring algorithms that significantly reduce the management overhead. The performance of the distributed monitoring algorithms is tested in both simulation testbed and real network, and compared with a centralized algorithm. The experimental results show that the distributed algorithm performs much better than the centralized algorithm.