Exploring botnet evolution via multidimensional models and visualisation

Abstract

A botnet is a program designed to perform a specific task using multiple computers connected in a network. In this paper we will focus on botnets being used to distribute malicious programs. In the real world, botnets have been shown to exhibit more aggressive and sophisticated behaviour than traditional malware. Botnets are used to infect computer networks and hence their success depends on the properties of the networks. We observe the behaviour of mathematical models used to describe botnets when botnet parameters are varied to understand if such variation is beneficial to their spread. We also introduce novel models for depicting botnet behaviour using master equations. These models, unlike previous ones, address nodes of distinct categories in a network as a sequence of probability distributions rather than a value at each time interval. We also contribute visualisations for these models. This paper is a substantial expansion of unpublished work the first author performed while on a Nuffield student research placement, with the second author the project supervisor.