ONR MURI area: High Confidence Real-Time Misuse and Anomaly Detection Intrusion and Anomaly Detection in Network Traffic Streams: Checking and Machine Learning Approaches ONR MURI area: High Confidence Real-Time Misuse and Anomaly Detection
Framework and System Architecture for Anomaly and Intrusion Detection Sampath Kannan Insup Lee Oleg Sokolsky Wenke Lee Diana Spears William Spears Linda Zhao
Overview Our approach is based on integration of a variety of anomaly and intrusion detection techniques A uniform mechanism and architecture is needed to support the integration Requirements: Flexibility Transparency Efficiency
MaC-based IDS
Background: MaC system MaC has been designed for run-time verification of software systems Main features: Checker decoupled from the system Event recognizer extracts relevant events from input stream Impact on reduced checking overhead
Background: MaC architecture
MaC extensions for IDS Multiple specification languages Dynamic property adjustment Checking of probabilistic properties
Integration architecture Unsupervised learner Cluster identification routines provide new detection rules Supervised learner Logistic regression modeling Tree-based algorithms Support vector machines
Integration architecture