An Efficient Approach in Protection of Information Security via Collaborative Inference Detection

This paper examines a wide variety of DoS and scanning attacks and shows that several categories (bandwidth based, claim-and-hold, port-scanning) can be scalably detected. In addition to existing approaches for scalable attack detection, the proposed innovative data structure called Fuzzy partial completion filter (FPCF) that can detect claim-and-hold attacks scalable in the network with the introduction of membership and non membership degree of attack prediction. This paper analyzes FPCF both analytically and using experiments on real network traces to demonstrate FPCF can be tuned to achieve extremely low false positive and false negative probabilities. The experimental result shows that the proposed method performs better than the existing partial completing filter.