Mitigating insider threat on database integrity

Abstract

We have developed a model to predict and prevent potential damage caused by malicious transactions in a database system. The model consists of a number of rules sets that constrain the relationships among data items and transactions. It uses a graph called Predictive Dependency Graph to determine data flow patterns among data items. The model offers a mechanism to monitor suspicious insiders activities and potential harm to the database. Through simulation we have tested the effectiveness of the model. The results show the effectiveness of the proposed model in predicting damage that can occur by malicious transactions. © Springer-Verlag 2012.