Abstract:
Security concerns have been raised by Web services providers and consumers since Web services are vulnerable to various security attacks including counterfeiting, disclosure, tampering, disruption, and breach of information. In particular, Web services can be vulnerable if the schemas of the input data are not strong, giving way to security attacks like command injection and denial of service. This research proposes an initial assessment of security attack risks for Web services. The assessment begins with an analysis of the input data schemas that are described in the service WSDL document to determine if they are unconstrained and at risk of command injection and denial of service attacks. Then we determine if such a risk can be mitigated by making use of semantic information that is annotated to the input data elements within the WSDL. If the semantic annotation is stronger than the schema elements themselves, we refer to the case of weak interface design in which a redesign of the service interface with stronger schemas should help reduce attack risks. We also propose a risk assessment model for determining quantitatively the attack risk level of a Web service to guide the provider when considering schema hardening and the consumer when selecting between different services.