Toward generating a DoS and scan statistical network traffic metrics for building intrusion detection solution based on machine and deep learning: I-Sec-IDS datasets.

We propose a Denial of Service (DoS) and scan statistical network traffic metrics datasets to build upon Intrusion Detection System (IDS) solutions built on Machine and Deep Learning (MDL) methodologies. We generate the datasets in VirtualBox environment. Two guests are involved and configured to perform the aforementioned attacks for collecting the network traffic. The first one is a Kali Linux VirtualBox machine that executes the DoS and scan attacks against the second one guest, a Microsoft 10. The host machine captures the exchanged network traffic between two guests via Wireshark and saves it in PCAP files. We extract the Canadian Institute of Cybersecurity (CIC)' FlowMeter metrics in Comma-separated Values (CSV) format to label our generated statistical network traffic metrics. Thus, this paper produces the DoS and scan statistical network traffic metrics datasets, cleansed up them before to be free available. To conclude, this work realises a first training datasets estimation to design an IDS solutions, based on MDL techniques, upon our datasets. [ABSTRACT FROM AUTHOR]