Would an Intrusion Detection System Perform Alike with the Change of the Number of Mobile Nodes? An Experimental Evaluation

Intrusion Detection Systems (IDSs) have recently been introduced to protect MANETs from malicious attacks that can reduce their performance. The performance of the IDSs proposed in the literature have been evaluated under certain network conditions such as the change of mobility speed, simulation time, mobility models, etc. However, none of these IDSs have been evaluated across the variation of mobile nodes (MNs). The increase of mobility nodes may directly impact the performance of the network and, consequently, the accuracy of the IDS. When the number of mobiles nodes goes up, intruders have better opportunity to spread over the existing paths and can trick the IDS. In this paper, we evaluate the effect of the increase/decrease of mobile nodes on the performance of IDSs. The IDS chosen in this paper is the Adaptive Acknowledgment (AACK). Our experiments are accomplished in NS2 and configured in a way all parameters set to be fixed during the entire simulation, except the number of mobile nodes (cooperative and malicious), which repeatedly kept increasing. Experimental results demonstrate the cases in which the performance of AACK is either improved or degraded when the number of MNs changes.