Researchers propose Katz centrality for securing critical networks
New paper shows how Katz centrality can predict and prevent stealthy cyber attacks on networked systems
Researchers Anh Tung Nguyen, Sribalaji C. Anand, and André M. H. Teixeira have published a novel security framework for protecting networked control systems against stealthy false data injection (FDI) attacks. The team proposes using Katz centrality—a graph theory metric that measures node influence based on both direct and indirect connections—to optimize security monitoring allocation.
The paper demonstrates that the worst-case network performance loss from stealthy attacks can be upper-bounded by a tractable semi-definite programming (SDP) problem. More importantly, they establish a relationship between this SDP problem and Katz centrality under sufficient conditions, reducing the security optimization to a network-size-independent problem. This enables defenders to prioritize monitoring of critical nodes without solving computationally intensive optimization problems in real-time scenarios.
- Proposes using Katz centrality to identify critical nodes in networked systems for security allocation
- Converts worst-case security optimization into a network-size-independent problem via SDP formulation
- Validated on Erdos-Renyi random graphs through extensive simulations
Why It Matters
Provides a computationally efficient way to protect critical infrastructure like power grids and industrial control systems from stealthy cyber attacks