Foundations for Network Resilience Against Cascading Failures - Modeling Fundamental Limits and Robust Network Architecture

Project: Research project

Description

The proposed work will tackle fundamental network research challenges in combating the attacks from Weapons of Mass Destruction (WMD). It will support the DTRA Initiative, Thrust 2, to advance the state of network science and the body of knowledge for combating WMD disruption and cascading failures. More specifically, aiming to preserve essential network capabilities in the nations cyber-physical infrastructures, this research project is focused on fundamental network theories and novel algorithms/protocols, for modeling, detecting, and stopping cascading failures due to WMD events, and for supporting network forensics, failure isolation, and network restoration. Based on the insight that cascading failures are spatially clustered, temporally correlated, and inter-dependent across different networks, we view a cyberenabled physical network as an overlay/underlay interplay system. We will develop interdependent Markov random field models and dependence graph models to quantify the spatial-temporal correlation structure of cascading failures and to capture the functional interactions among physical and communication components across multiple networks. Building on these mathematical models for cascading failures, we then study network inference algorithms to detect the presence of cascading failures in early stages, estimate their damages, and stop their propagation. Robust network architecture will be investigated in depth to achieve the optimal allocation of interconnecting links and improve network seal-healing; and cross-network optimization will be investigated to further enable the design for resilience to cascading failures. The PIs track records, together with an innovative agenda for basic research, provide confidence that the proposed work will not only advance the field of network adaptability from WMD disruptions, but also contribute to fundamental theories and methodologies useful for social-tech networks and other DoD-operated networks. The proposal also contains an education plan for training students and postdocs.
StatusActive
Effective start/end date1/7/1312/6/19

Funding

  • DOD: Defense Threat Reduction Agency (DTRA): $1,235,844.00

Fingerprint

Network architecture
Circuit theory
Restoration
Seals
Education
Mathematical models
Students
Network protocols
Communication