Enterprise Data Integration towards Web Service Personalization

Service personalization and automated service composition are two active research focii in web services domain. Currently, the systems developed for web service personalization require users to furnish fine-grained information about their request during the interaction process. In order to relief the burden of active role played by the users whenever personalized web services are requested, we adopt agent-based supply chain techniques. Supply chain management has been successfully used in product built-up, where different stages including ordering, planning, supply, and production are networked together towards manufacturing preferred products for customers. We present in this paper how the supply stage in the technique is effectively maintained in our system. Query requests are sent to external heterogeneous data sources - with structured, semi-structured and unstructured format - where data for the personalization reside and are elicited from as the need arises through a brokering service. We evaluated the rate of accessing these different data sources and made a comparison between them. The result shows that the data read directly via storage in a structured database table are accessed more quickly compare to those accessed from other repositories.