Real-time GPU-based timing channel detection using entropy

摘要: As line rates continue to grow, network security applications such as covert timing channel (CTC) detection must utilize new techniques for processing flows in order protect critical enterprise networks. GPU-based packet provides one means of scaling the CTCs and other anomalies flows. In this paper, we implement a tool, capable detecting model-based channels (MBCTCs). The GPU's ability process large number packets parallel enables more complex tests, corrected conditional entropy (CCE) test—a modified version measurement, which has variety outside detection. our experiments, evaluate CCE test's true false positive rates, well time required perform test on GPU. Our results demonstrate that GPU can be applied successfully real-time CTC at near 10 Gbps with high accuracy.