Dr. Antonakakis’s presentation, “From Throw-Away Traffic to Bots: Detecting the Rise of DGA-Based Malware,” is based on a paper written by researchers at Damballa Labs, Georgia Institute of Technology and the University of Georgia. The paper can be viewed here: https://www.usenix.org/conference/usenixsecurity12/throw-away-traffic-bots-detecting-rise-dga-based-malware.
The session will introduce a new cyber threat detection technique named Pleiades, which is designed to detect criminal malware threats that use domain generation algorithm (DGA)-based techniques to establish command-and-control communications and completely evade detection by blacklists and signature-based systems. The Pleiades innovation, which is already deployed in Damballa customer installations, also eliminates the need to reverse engineer malware in an attempt to ‘decode’ the DGAs, a method which is often unsuccessful or unreliable. The new technology recently enabled Damballa customers to detect the Flashback malware, which ultimately infected more than 600,000 Macintosh devices, weeks before the malware was first discovered and announced by the security community.
Pleiades uses a combination of clustering and classification algorithms. The clustering algorithms cluster domains based on the similarity in the make-up of domain names, as well as the groups of end point devices that queried the domains. The classification algorithm is used to assign the generated clusters to models of known DGAs. If a cluster cannot be assigned to a known model, then a new model is produced, indicating a new DGA variant or family.
During the session, Dr. Antonakakis will report the discovery of twelve DGAs, half of which are variants of known malware DGAs, and the other half are new DGAs that have never before been reported.
This research follows closely a discovery which Damballa unveiled earlier this year on advanced evasion techniques using DGAs by six crimeware families to carry out global cyber attacks. Without having to reverse engineer malware or 'decode' the DGA algorithm, Damballa Labs automatically detected and modeled DGA behavior by using this patent-pending machine learning technology. The report is titled “DGAs in the Hands of Cyber-Criminals - Examining the State of the Art in Malware Evasion Techniques” and can be viewed here:
The conference takes place August 8-10 at the Hyatt Regency Bellevue, in Bellevue, WA. The full agenda can be found at: https://www.usenix.org/conference/usenixsecurity12/glance.
About Damballa - Damballa is a leading provider of advanced threat protection solutions for corporate, telecommunications and Internet service provider networks. Damballa provides the only network security solution that detects both criminal command-and-control (C&C) behaviors and inbound malware; automatically correlating all evidence of criminal behavior to uncover hidden infections and terminate the criminal activity. Patent-pending solutions from Damballa protect networks with any type of server or endpoint device including PCs, Macs, Unix, smartphones, mobile and embedded systems. Damballa protects more than 200 million endpoints worldwide at mid-size and large enterprises in every major market. http://www.damballa.com