Investigation on operating systems identification by means of fractal geometry

Abstract

This work demonstrates a novel way on how to identify the operating systems and networking devices working with TCP/IP stack on the basis of differences in their pseudorandom number generators. Data from pseudorandom number generators of various OS are used to be visualized in its randomness in 3D space, using reconstruction method from deterministic chaos. Then fractal geometry is applied to measure fractal dimension of observed 3D clouds and compared with Euclid objects and among themselves. Interesting fractal properties are revealed and discussed in this article as well as sketch of possible OS identification based on fractal geometry and neural networks.