Covert Channels Detection with Supported Vector Machine and Hyperbolic Hopfield Neural Network

Abstract

A mechanism that is intended to expose information against a security violation in a network is the use of network covert channel and it is difficult to detect information about data loss like location of loss using network covert channel. To identify the covert channel were the data pattern missing over the sharing of resources in networks. Several mechanisms are used to identify a large variation of covert channels. However, those mechanisms have more limitation like speed of detection, detection accuracy etc. In this paper, a new machine learning approaches called “Support Vector Machine and Hyperbolic Hopfield Neural Network” to overcome the drawbacks of existing methods. This approach is supported to classifying the different covert channels with data packets which is shared in networks and its supports to identifying the location of data loss or data pattern mismatched. Finally, the proposed methods properly detected covert channels with high accuracy and less detection high speed shared a network resources in effective manner.