Identifying control and management plane poison message failure by K-nearest neighbor method

Poison message failure is a mechanism that has been responsible for large-scale failures in both telecommunications and IP networks. The poison message failure can propagate in the network and cause unstable network. In this paper, we apply machine learning, data mining technique in network fault management area. We use k-nearest neighbor method to identify the poison message failure. Also we integrate the k-nearest neighbor method with message filtering approach. We also propose a "probabilistic" k-nearest neighbor method that outputs a probability distribution (rather than the identity) of the poison message. Through extensive simulations, we show that k-nearest neighbor method is very effective in identifying the responsible message type.