Mining data relationships for database damage assessment in a post information warfare scenario

摘要

After the detection of a cyber attack on a database system, the intrusion response team of any organization needs to know the damage profile immediately in order to design an appropriate response strategy. Unfortunately obtaining the precise damage status can take up to hours even days. This is because existing approaches to database damage assessment involve significant amount of work including scanning the log file or other auxiliary data structures. Our approach concentrates on making an estimated damage profile as soon as possible. This model is based exclusively on a priori knowledge of data relationships mined during normal database operation phase. This knowledge can be used during damage assessment phase for faster damage assessment.