MASRR: Multi-Agent System for network Resource Reliability

Intelligent, automated solution to network management Autonomously diagnoses and responds to problems Discovers and corrects network misconfigurations Detects attacks and intrusions early to minimize damage Functions robustly, even when part of the network has failed or is under attack Adapts to changing networks and usage Works on heterogeneous networks Manages networks of any size without excessive resource consumption Today's economy depends on reliable computer and communications network services. However, networks have grown in complexity, and reliability is harder than ever to manage. Networks are composed of different computer platforms, legacy systems, and proprietary hardware interfaces. Even the experienced administrator wages a battle to stay current with the changing protocols, applications, and security alerts. While generally robust, networks remain vulnerable to hardware and software failures, attacks, and misuses of seemingly unlimited kinds. We have seen networks brought down by faulty switches, e-mail viruses, and denial of service (DoS) attacks. The growing complexity of networks and our reliance on them is exacerbated by a shortage of expert administrators, and calls for an intelligent, automated solution to network management. To that end, SHAI introduces Multi-Agent System for network Resource Reliability (MASRR). Using models of normal behavior of network components and users, as well as knowledge of known abnormal or problematic behaviors, software agents diagnose failures, misconfigurations and attacks, as well as detect and respond to unknown anomalies. Prognostic simulation allows agents to actually predict faults and take actions to report, minimize, and contain the damage. Background Project Focus Innovations Technical Reports and Papers Presentations Artificial Intelligence (AI) technologies  MASRR uses these AI techniques to achieve its goals: agents case based reasoning datamining machine learning time series prediction   Description of how these technologies are applied in MASRR For more information on MASRR, please contact the project manager, Lynn Jones. Other Computer Security Projects at Stottler Henke

Project Status Results running MASRR in SSFNet simulator Technical Reports & Papers Towards Decentralized Network Management and Reliability, in proceedings of The 2002 International Conference on Security and Management, (SAM'02, Las Vegas, June 24-27). [pdf] Handling router misconfiguration [html] Cooperative response to an unknown anomaly [html] Change and Anomaly Detection (ChAD) [html] Presentations Towards Decentralized Network Management and Reliability, (SAM'02, paper available above). [power point] overview given at Network Modeling and Simulation Workshop at Dartmouth College, July 2002. [power point] Interactive Use of SSFNet for Creating a Testbed. [power point] ChAD: Change and Anomaly Detection, 12/01 [power point]