Spatiotemporal Cyberspace Situation Awareness Mechanism for Backbone Networks

As the future key research direction of network management, cyberspace situation awareness (CSA) can provide a macroscopic view of network state and build resilience into network control. A spatiotemporal CSA mechanism based on information fusion (IF) for backbone network is proposed, in which the situation assessment and prediction are included in a closed-control-loop. The paper establishes an IF system that adopts probabilistic graphical methods for situation assessment using both Hidden Markov Model (HMM) for temporal modeling and Hidden Conditional Random Fields (HCRF) for spatial modeling, and the Radial Basis Function Neural Network (RBFNN) trained by Artificial Bee Colony (ABC) is taken for situation prediction. Since we aim to minimize the position selecting deviation of traffic sensor deployment in IF system, it is achieved by formulating the problem as a constrained convex optimization problem subject to information leakage guarantee. The experiment results verify the IF-based mechanism on real backbone network traffic dataset and simulate the CSA system with performance evaluation and comparison.