Development and Evaluation of Defaced Sites Automatic Detection System DICE

Recently, the falsification attack to bury an illegal script under the Web site by using SQL injection becomes a problem. Authors previously proposed the detection method by the algorithm made based on the site title actually obtained through the survey of a falsification site and an illegal script and the analysis data of the description pattern of the script. In this study, we improved the method to obtain higher detection rate by Hayashi´s quantification theory type 2. We also implemented the old and new methods as gDefaced Sites Automatic Detection System DICEh on a proxy server program named Squid. This paper reports the method, the implemented program and the results of functional experiment and performance experiment.