Visual Characterization of the Dynamics for Network Data Segments (Port 110) with Nonlinear and Computational Techniques

Abstract

The collective behavior of network data segments corresponding to port 110 (e-mail) was visualized and analyzed with a set of nonlinear and computational parameters such as: Lyapunov exponent, fractal dimension, recurrence, determinism, grammatical rules, and lempel-Ziv complexity. Common methods for network data analysis are based mainly in descriptive statistics, although useful and widely accepted they provide us with a global and partial view of the network dynamics. In this work we applied a series of analytical techniques from nonlinear and computational origin in order to elaborate a more complete view of the network dynamics for data segments from port 110. These techniques preserve information of local and global nature and their combination allows a more complete view about the collective dynamics of this network traffic. The visual representation with visual recurrence analysis and grammatical rules is useful for the detection of changes and events in the network dynamics.