Quantitative analysis of information leakage in service-oriented architecture-based Web services

Abstract

Purpose Any computing architecture cannot be designed with complete confidentiality. As a result, at any point, it may leak the information. So, it is important to decide leakage threshold in any computing architecture. To prevent leakage more than the predefined threshold, quantitative analysis is helpful. This paper aims to provide a method to quantify information leakage in service-oriented architecture (SOA)-based Web services. Design/methodology/approach To visualize the dynamic binding of SOA components, first, the orchestration of components is modeled. The modeling helps to information-theoretically quantify information leakage in SOA-based Web services. Then, the paper considers the non-interference policy in a global way to quantify information leakage. It considers not only variables which interfere with security sensitive content but also other architectural parameters to quantify leakage in Web services. To illustrate the attacker’s ability, a strong threat model has been proposed in the paper. Findings The paper finds that information leakage can be quantified in SOA-based Web services by considering parameters that interfere with security sensitive content and information theory. A hypothetical case study scenario of flight ticket booking Web services has been considered in the present paper in which leakage of 18.89 per cent information is calculated. Originality/value The paper shows that it is practically possible to quantify information leakage in SOA-based Web services. While modeling the SOA-based Web services, it will be of help to architects to identify parameters which may cause the leakage of secret contents.