CONFIDANT: Collaborative Object Notification Framework for Insider Defense using Autonomous Network Transactions

File Integrity Analyzers serve as a component of an Intrusion Detection environment by performing filesystem inspections to verify the content of security-critical files in order to detect suspicious modification. Existing file integrity frameworks exhibit single point-of-failure exposures. The Collaborative Object Notification Framework for Insider Defense using Autonomous Network Transactions (CONFIDANT) framework aims at fail-safe and trusted detection of unauthorized modifications to executable, data, and configuration files. In this paper, an IDS architecture taxonomy is proposed to classify and compare CONFIDANT with existing frameworks. The CONFIDANT file integrity verification framework is then defined and evaluated. CONFIDANT utilizes three echelons of agent interaction and four autonomous behaviors. Sensor agents in the lowest echelon comprise the sensor level to generate an assured report to companion agents of computed MD5 file digests. At the control level, beacon agents ...

• Type: Journal
• Year: 2006
• Where: AAMAS
• Authors: Adam J. Rocke, Ronald F. DeMara
• Keywords: AAMAS 2006, CONFIDANT, CONFIDANT File Integrity, File Integrity, Intelligent Agents