Packet filtering by artificial neural network

Abstract

Efficient monitoring of the network is very important for an achievement of a security of today networks. The still growing speed of the links and the complexness of monitoring applications' requests have showed some borders of mostly used methods for monitoring. The process of packet classification should be speeding up as much as possible. As a possible approach, an artificial neural network (ANN) could be used for packet filtering. The performance of the artificial neural network was validated by software implementation of ANN for given network configuration. The principles of artificial neural networks and the possibility of using artificial neural network in a computer network are presented in the article. Created training sets represent the information for firewall to enable or disable packets. The number of neurons in the artificial neural network and the number of the hidden layers are optimized based on the results in the simulation.