A Parallel Approach to PCA Based Malicious Activity Detection in Distributed Honeypot Data

Model order selection (MOS) schemes, which are frequently employed inseveral signal processing applications, are shown to be effective tools for the detectionof malicious activities in honeypot data. In this paper, we extend previous results byproposing an efficient and parallel MOS method for blind automatic malicious activitydetection in distributed honeypots. Our proposed scheme does not require any previousinformation on attacks or human intervention. We model network traffic data as signalsand noise and then apply modified signal processing methods. However, differently fromthe previous centralized solutions, we propose that the data colected by each honeypotnode be processed by nodes in a cluster (that may consist of the collection nodesthemselves) and then grouped to obtain the final results. This is achieved by having eachnode locally compute the Eigenvalue Decomposition (EVD) to its own sample correlationmatrix (obtained from the honeypot data) and transmit the resulting eigenvalues to acentral node, where the global eigenvalues and final model order are computed. Themodel order computed from the global eigenvalues through RADOI represents the numberof malicious activities detected in the analysed data. The feasibility of the proposedapproach is demonstrated through simulation experiments