Description

•                                        Abstract—Data as a Service (DaaS) builds  on service-oriented technologies to enable fast  access to data  resources on the  Web. However,   this  paradigm raises  several  new  privacy  concerns that  traditional  privacy  models do  not  handle.  In  addition,  DaaS composition may  reveal  privacy-sensitive information.  In this  paper, we  propose a  formal privacy  model  in order  to extend DaaS descriptions with privacy capabilities. The privacy model allows a service to define  a privacy policy and  a set of privacy requirements. We also  propose a privacy-preserving DaaS composition approach allowing to verify the compatibility between privacy requirements and  policies  in DaaS composition. We propose a negotiation mechanism that  makes it possible to dynamically  reconcile  the privacy capabilities of services when incompatibilities  arise  in a composition. We validate  the applicability of our proposal through  a prototype implementation and a set of experiments.