| Service composition is widely used in enterprises to combine and reuse existing services in order to realize some set of requirements not fulfilled by a single service. In this context, the service composition process is performed by professional service developers, which have technical knowledge and skills in the composition process. We foresee that non-professional users can also benefit from service composition, namely as a mechanism to simplify and personalise service delivery. We define such a personalized service creation that takes user-specific requirements and properties into account User-centric service composition . The support of user-centric service composition, where the users are mainly non-professional users, is a difficult undertaking. Mechanisms that shield users from the details and technicalities of the service composition process are required. Furthermore, non-professional users are not homogeneous; they may have different knowledge of the domains where they are seeking services or have different skills on the systems that support the composition process. To address user-centric service composition we are investigating and developing mechanisms to support/automate the different phases of the service composition life-cycle. Namely, we are developing algorithms and techniques that use semantic web technologies to perform automatic service discovery and composition. We are also investigating how we can make the composition process user-centric, i.e., consider the properties of the specific user that is being supported. To address this we are designing and developing an architecture that can be adapted to the user that is being supported in a given moment. The user properties we consider in this adaptation process are: application domain knowledge, technical skills and how the user interacts with the system. Based on this knowledge, the supporting architecture adapts to better suit the user being supported. New users and new supporting functionality can be accommodated in the architecture, granting evolvability and extensibility. This work is being implemented and experimented in the DynamiCoS (Dynamic Composition of Services) framework. |
