Defending Web Applications Against JavaScript Worms on Core Network of Cloud Platforms

Abstract

In this article, the authors suggest a cloud-hosted defensive model that defends the cloud-hosted web applications against the injection of JavaScript (JS) worm injection vulnerabilities. The model is categorized into two modes, namely, training mode and detection mode. The earlier one involves calculating and storing the Script Property Functions of JS code in the Script Property Function Repository. The robustness of our framework lies in the method of Script Property Function comparison before executing the browser environment-independent sanitization of JS code. It also involves extracting JS code from the HTTP response and comparing it with the JS code extracted from Uniform Resource Identifier (URI) links embedded in HTTP request. The other mode compares the JS Script Property Functions computed in two different modes. Any variance observed in both these calculated Script Property Functions sets results in consequent sanitization of the JS code. In this work, we have developed our prototype on the environmental set-up of ICAN cloud simulator and its settings were integrated by creating the infrastructure set-up of various virtual machines on this simulator. The proposed scheme is proficient enough to detect and eliminate the JS worms from the tested web applications with high precision rate, tolerable rate of False Negatives (FNs) and False Positives (FPs).