A Hybrid Architecture for Database Intrusion Preventer

Abstract

Database management systems come with several security mechanisms such as access control, encryption and auditing. These mechanisms are very important to protect databases against various threats. However, such mechanisms may not be sufficient in dealing with database intrusion. Therefore, the prevention of any intrusion attempts as well as detecting them pose an important research issue. Despite the proposal of many techniques previously, the design and implementation of database reliable intrusion detection or prevention systems remains a substantial demand and a vital research topic. In this paper, a Hybrid Architecture for Database Intrusion Preventer (HyDBIP) has been proposed. The proposed system comprises of Signature-based, Anomaly-based and Anomaly Query Classifier (AQC) models work together to complement each other.