RPAD: Rule based Pattern Discovery for Input Type Validation Vulnerabilities Detection & Prevention of HTTP Requests

摘要

The internet access by web browsers is most vulnerable, since the browsers itself can adapt as attacking tool. The traditional way of dealing the influence of these vulnerabilities such as SQL Injections and XSS is the code verification by syntax analyzers. Since the syntax analyzers deployed and executed in server environment, the process overhead on servers is the major constraint observed, which is due to the online verification of dynamic SQL statements generated by web enabled applications and injected into server-side applications. The other major constraint of these syntax analyzers programming language dependency. In order to this here in this paper we proposed a Rule based Pattern Discovery (RPAD) for Input Type Validation Vulnerabilities Detection and Prevention of HTTP Requests. The core objective of the RPAD is to execute as network level IDS and not to rely on syntax analyzers, hence the limits such as programming language dependency and server level process overhead observed in existing benchmarking models are least significant for RPAD. The other competency of RPAD is minimal sanity checks. The experimental study was conducted on several benchmarking CVE entries published by NIST. The combination of 2783 attack patterns extracted from CVE entries of the NIST and 512 normal patterns extracted from 7 real time web applications were used to evince the performance of the RPAD. The empirical study evinced that the RPAD prediction accuracy is around 93%.