Professor
Sungdeok Cha
- scha@korea.ac.kr
- Homepage
- http://dependable.korea.ac.kr
- Faculty Office
- Science Library 617B
Fields
-
Dependable Software Laboratory (DSL) at Korea University, led by Prof. Sungdeok (Steve) Cha, conducts research on software engineering and computer security techniques that are essential in building safe and reliable systems. Building on the research results accomplished at the Dependable Software Laboratory at Korea Advanced Institute of Science and Technology (KAIST), DSL at Korea University is working on the following topics: Software embedded in medical devices. Examples include artificial heart, robotic surgery system, etc. We’re currently analyzing software controlling artificial heart developed by Korea Artificial Organ Center (KAOC). We plan to extend our analysis to software running on robotic surgery system in the near future. Real-Time Anomaly Detection in Web Servers. With generous support from Microsoft Research Asia, we analyzed web logs collected in a day, nearly 250GB in size and has been made anonymous to preserve privacy, and classified features that effectively characterize various web robots. We’re currently developing a system, tentatively named CAV (Composite Attribute Vector), which visualizes activities of web robots in real-time. Features unique to CAV include 1) Selective display of only relevant information to avoid information overflow so that the system is scalable enough to use in large web servers; and 2) Its ability to explain why specific session was classified as having been initiated by a web robot.