Search

The surge in the number of services being provided off cloud-based servers in recent times, has led to a concern in the security of these platforms. It is thus highly important to have in place systems which can efficiently recognize and alert the system administrators about malicious activities on the cloud platform. Further, it is desirable to know the kind of attack being made on the cloud server to better understand and mitigate the threats. The usage of Artificial Neural Networks(ANN) has been found in Intrusion Detection Systems(IDS). To bring these systems closer to the biological model of the neural system, in this paper, the authors have deployed Spiking Neural Networks (SNN) to create a highly efficient IDS. SNNs deploy spiking neurons which mimic the biological neuron more closely than the artificial neuron model which is an overly simplified mathematical model. In this paper, the authors propose a hybrid SNN-ANN model which achieves a high degree of accuracy in identifying and classifying malicious connections on the network into their type of attack using the NSL-KDD dataset.