SQL Injection Attack Vulnerability Mitigation Method Using Keyword Classification and ANN-based Query Reconstruction Scheme

Abstract

Today, in the digital information era, the Internet and AI technology have changed the lives of people greatly. Traditional cyber-attack detection is no longer available in today’s complex network environment, and machine learning technology using artificial intelligence has started to play an important role in the network security field. In order to enhance the security of the web system, it is very urgent to establish a strong security policy to prevent the web-front-end script attack, such as cross-site scripting attack (XSS), and to prevent the web-server script attack, i.e., SQL injection attack. SQL injection attack is one of the most common application-layer attack techniques used today. Various methods to prevent modern SQL injection attacks have been investigated and introduced to ensure the safety and reliability of the system. SQL injection attacks occur due to insufficient testing of user queries at the client, and to prevent this, we have to apply a method that performs an accurate check of the input fields. The main challenge in preventing SQL injection attacks is to reduce the execution overhead and increase the detection rate for different attack types. We propose an attack detection method based on an ANN that enhances the detection rate against SQL injection attacks and provides maximum service for authenticated users, and we evaluate the security and performance of the system.