Security in Web Applications: A Comparative Analysis of Key SQL Injection Detection Techniques

Over the years, technological advances have driven massive proliferation of web systems and businesses have harbored a seemingly insatiable need for Internet systems and services. Whilst data is considered as a key asset to businesses and that their security is of extreme importance, there has been growing cybersecurity threats faced by web systems. One of the key attacks that web applications are vulnerable to is SQL injection (SQLi) attacks and successful attacks can reveal sensitive information to attackers or even deface web systems. As part of SQLi defence strategy, effective detection of SQLi attacks is important. Even though different techniques have been devised over the years to detect SQLi attacks, limited work has been undertaken to review and compare the effectiveness of these detection techniques. As such, in order to address this gap in literature, this paper performs a review and comparative analysis of the different SQLi detection techniques, with the aim to detect SQLi attacks in an effective manner and enhance the security of web applications. As part of the investigation, seven SQLi detection techniques including machine learning based detection are reviewed and their effectiveness against different types of SQLi attacks are compared. Results identified positive tainting and adoption of machine learning among the most effective techniques and stored procedures based SQLi as the most challenging attack to detect.