Security Analytics for Defeating Automated Internet-Scale Threats

Friday, February 24, 2017 noon to 1 p.m.

Aziz Mohaisen

Assistant Professor, University of Buffalo.

Billions of devices are connected to the Internet today, significantly changing the threat landscape by lending adversaries unprecedented resources to launch automated attacks, and requiring new threat analysis and defenses. In this talk, I will argue that big data analytics can play an important role in securing the Internet, and exemplify my argument with applicaƟons to distributed denial of service (DDoS), malware analysis, and massively multiplayer online role‐playing game (MMORPG) bot detection. First, I will present an analytical view of 50,000 unique and verified DDoS attacks on services on the Internet. I will show how adversaries’ spatiotemporal traits follow predictable patterns, consecutive attacks follow certain patterns allowing prediction of future threat, and attackers are highly collaborative. Second, I will show how big data analytics are applied to malware analysis and software behavior profiling, and demonstrate optimizations to scale such analytics. Third, I will discuss an analytics framework for game bot detection in MMORPG using self‐similarity of user behavior. By applying this framework to three large online games, I demonstrate how this analytics approach can be used to extract general features of behavior and effectively detect game bots in practice. I will conclude by highlighting my vision of how this analytics approach can be applied to realize effective and proactive defenses, and extended for other applications.

Hosted by: Dr. Gary Leavens

http://www.eecs.ucf.edu/seminar_flyers/24February2017.pdf

Read More

Location:

Teaching Academy: 117

Contact:

Dr. Gary Leavens Gary.Leavens@ucf.edu

Calendar:

CS/CRCV Seminars

Category:

Speaker/Lecture/Seminar

Tags:

Computer Science