Influence of network payload and traffic models on the detection performance of AIS

We investigate the influence of the network traffic payload, using 50 concurrent connections, with a Poisson distributed packet injection model, on the detection performance of our Artificial Immune System (AIS). We compare the detection performance to priorly gained results which were based on a smaller scenario. We conclude that the Poisson traffic model had again no negative impact on the detection performance. We also conclude that a higher network payload has no negative impact on the detection performance. Additionally a statistically significant difference in the detection performance between CBR and Poisson could be observed for a high network payload.