A method of classifying multimedia traffic flows based on neural network

A new method of classifying multimedia traffic flows was proposed for the next generation Internet (NGI). By analyzing the time level correlation of traffic flows, a model of competitive artificial neural network was designed to identify and classify different traffic flows and modulate them which can help to marking the Ipv6 flow label. Simulation result shows that the method can increase the precision of classification, and has little impact on the throughput of border gateway, has actual value.