Requirements for Training and Evaluation Dataset of Network and Host Intrusion Detection System

摘要: In the cyber domain, situational awareness of critical assets is extremely important. For achieving comprehensive awareness, accurate sensor information required. An important branch sensors are Intrusion Detection Systems (IDS), especially anomaly based intrusion detection systems applying artificial intelligence or machine learning for detection. This millennium has seen transformation industries due to developments in data modelling methods. The most crucial bottleneck IDS absence publicly available datasets compliant modern equipment, system design standards and threat landscape. predominant dataset, KDD Cup 1999, still actively used research despite expressed criticism. Other, more recent datasets, tend record only either from perimeters testbed environment’s network traffic effects that malware on a single host machine. Our study focuses forming set requirements holistic Network Host System (NHIDS) dataset by reviewing existing studied within field modelling. As result, state-of-the-art NHIDS presented be utilised development intelligence.