Achin Jain and Vanita Jain
Bharati Vidyapeeth College of Engineering, India

In the present scenario, protection of websites from web-based attacks is a great challenge due to the bad intention of the malicious user over the Internet. Researchers are trying to find the optimum solution to prevent these web attack activities. There are several techniques available to prevent the web attacks from happening like firewalls, but most of the firewall is not designed to prevent the attack against the websites. Moreover, firewalls mostly work on signature-based detection method. In this paper, we have analyzed different anomaly-based detection methods for the detection of web-based attacks initiated by malicious users. Working of these methods is in a different direction to the signature-based detection method which only detects the web-based attacks for which a signature has been previously created. In this paper, we have introduced two methods: Attribute Length Method (ALM) and Attribute Character Distribution Method (ACDM) which is based on attribute values. Further, we have done the mathematical analysis of three different web attacks and compare their False Accept Rate (FAR) results for both the methods. Results analysis reveals that ALM is more efficient method than ACDM in the detection of web-based attacks

Web-Based Attacks, Anomaly Based Detection, Attribute Length Method, Attribute Character Distribution Method, False Accept Rate