SQL Injection Detection for Web Applications Based on Elastic-Pooling CNN

Abstract

An enterprise's data can be one of its most important assets and often critical to the firm's development and survival. SQL injection attack is ranked first in the top ten risks to network applications by the Open Web Application Security Project (OWASP). Its harmfulness, universality, and severe situation are self-evident. This paper presents a method of SQL injection detection based on Elastic-Pooling CNN (EP-CNN) and compares it with traditional detection methods. This method can output a fixed two-dimensional matrix without truncating data and effectively detects the SQL injection of web applications. Based on the irregular matching characteristics, it can identify new attacks and is harder to bypass.

Keywords

• Deep learning
• neural network
• CNN
• network security
• SQL injection