Modeling and measuring privacy risks in QoS Web services

The privacy issue for QoS Web services is considered. As more Web services are constructed and deployed, many of our personal needs may be served by more than one service providers regardless of their platforms and implementation technologies. Service selection is performed by comparing the QoS attributes of all candidate services. Releasing private data without due control may raise the risk of allowing potential adversaries to collect, reveal or utilize one's data in undesirable ways. We propose a QoS model that quantifies a user's privacy risk in order to make the service selection process manageable. The privacy risk is derived using the percentage of the private data set to be released, and user-defined context-oriented weights that quantify the potential damage of privacy leaking in a specific context. We have included privacy risk cost as one of the QoS parameters by extending the end-to-end QoS composition algorithms developed previously.