Traffic Monitoring and Malicious Detection Multidimensional PCAP Data Using Optimized LSTM RNN

Nowadays, the intrusion detection systems (IDSs) and network security assessments utilize the methodology of deep learning with several innovations like recurrent neural networks (RNN) and long short-term memory (LSTM) for classifying the malicious traffic. For satisfying the requirements of real-time analysis because of main delay of the flow-based data minimization, these state-of-the-art systems face enormous challenges. The flow-based minimization is the time required for specific flow of packet accumulation and then feature extraction. In case the detection of malicious traffic at the packet level is accomplished first, and then significant reduction of time for detection happens, this ensures the online real-time malicious traffic detection depends upon the technologies of deep learning as a promising one.