Application of Sliding Window Deep Learning for Intrusion Detection in Fog Computing

Fog computing is an extended concept of cloud computing or edge computing. It was introduced to reduces the overhead and existing issues of cloud such as latency, wastage of resources, mobility support, etc. As fog computing is decentralized architecture to process incoming network traffic. With increasing traffic, there is an increase in network threats. To handle such threats, Intrusion Detection Systems (IDSs) provide secure network traffic flow. Due to the limitation of resources over the fog network, it is required to design lightweight IDS. In this paper, to provide efficient detection of incoming network traffic, a deep learning technique is proposed and implemented with a sliding window that identifies the attacks over network. The result analysis was performed by in-creasing the window size and showed up its efficiency with other existing techniques. The result was evaluated over the NSL-KDD dataset. The testing scenario was performed over multi-classification of data packets. The highest accuracy of approx. 99% was achieved over the NSL-KDD dataset.