Flow classification schemes in traffic-based multilayer IP switching - comparison between conventional and neural approach

In this work we compare different methods for router workload deduction and flow classification schemes. Gigabit-routing solutions and different schemes of integrated switching and routing (Internet protocol (IP) switching) have emerged. Several flow classification schemes to be used in traffic-controled IP switching are presented and evaluated. As a new alternative, the ‘Learning Vector Quantization’ classifier is examined and found to be able to successfully classify Internet traffic flows. q 1998 Elsevier Science B.V.