Cross-Analysis of Botnet Victims: New Insights and Implications Conference Paper uri icon

abstract

  • In this paper, we analyze a large amount of infection data for three major botnets: Conficker, MegaD, and Srizbi. These botnets represent two distinct types of botnets in terms of the methods they use to recruit new victims. We propose the use of cross-analysis between these different types of botnets as well as between botnets of the same type in order to gain insights into the nature of their infection. In this analysis, we examine commonly-infected networks which appear to be extremely prone to malware infection. We provide an in-depth passive and active measurement study to have a fine-grained view of the similarities and differences for the two infection types. Based on our cross-analysis results, we further derive new implications and insights for defense. For example, we empirically show the promising power of cross-prediction of new unknown botnet victim networks using historic infection data of some known botnet that uses the same infection type with more than 80% accuracy. 2011 Springer-Verlag.

published proceedings

  • Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

author list (cited authors)

  • Shin, S., Lin, R., & Gu, G.

citation count

  • 10

complete list of authors

  • Shin, Seungwon||Lin, Raymond||Gu, Guofei

publication date

  • December 2011