On Resilience and Defense against Attack-Induced Failures in Cyber-Physical Systems
Abstract
Real-time cyber-physical systems are widely deployed in various computing and networking scenarios andapplications. Compared with conventional computer systems, cyber-physical systems feature critical real-timeinteractions between the physical infrastructures and computational devices to guarantee secure and reliable system operations, which are particularly important to mission-critical applications. For example, an on-board power plant is a critical component for any naval ship system and therefore must be closely monitored and tightly secured. One of the most representative cyber-physical systems today is the electric power system. A ship-board power system features the integration of a critical power generation and supply system that generates and transmits the power to all on-board electric devices, a computing, command and control network, and real-time quality-of-service (QoS) requirements. The system, by integrating computing and network devices into the physical power infrastructure, provides real-time monitoring and reliable control of the physical world. However, the system can still be potentially vulnerable to intentional cyber attacks. Malware or virus infection on computational devices connected to physical infrastructures can block or even falsify the observation results from monitoring functions of critical physical component operations, such as power generation components. Nonetheless, it is still unclear in the existing studies how the malware-induced attacks at the component level of the cyber domain can effectively block reliable interactions between the cyber and physical domains, leading to component and global failures in monitoring and controlling functions (e.g., power generation). This is particularly important for ship-board systems as power generation and supply components are critical for naval operations. This proposed project aims to address the timely issues in (i) characterizing attack-induced failures, (ii) modeling the system resilience using quantitative metrics, and (iii) providing the cyber defense approaches againstattackers with the intent to cause critical component-level or global failures in cyber-physical systems. The project is expected to provide fundamental techniques that can clearly quantify the impact of cyber attacks as well as the effectiveness of the best defense strategy to improve the resilience against attack-induced failures. The techniques can also potentially provide fundamental guidelines to security deployment in ship-board cyber-physical power systems with future naval applications.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2017
- Source ID
- N000141712109
Entities
People
- Zhuo Lu
Organizations
- Office of Naval Research
- United States Navy
- University of South Florida