PeANFIS-FARM Framework in Defending against Web Service Attacks

摘要

Internet-enabled Web Service (WS) applications, such as e-commerce, are facing eXtensible Markup Language (XML)-related security threats. However, network and host-based intrusion (ID) and prevention (IP) systems and Web Service Security (WSS) standards are inadequate in countering against these threats. This paper presents a framework to mitigate XML/SOAP attacks. Our framework comprises of two intelligent models: the policy-enhanced adaptive neuro-fuzzy inference system (PeANFIS) and fuzzy association rule mining (FARM) model. Performance evaluation of each model indicates detection rate of greater than 99% and false alarm rate of less than 1%. In this paper, we aim to help the security administrator to decide which model to implement depending on the context of the situation. We present rule-based cases as examples to guide design and implementation decisions. Our future work shall see the implementation of the PeANFIS-FARM framework on a wider scale and in cloud computing.