Data integration (DI) is usually used to combine a myriad of configurations, databases and applications. However, it has been implemented in a tactical and monolithic manner. Applications are tightly bound to their DI-specific logics, including creating, reading, updating and deleting (CRUD) data, data validation, transformation, and movement. Data Services are introduced to enable the insulation of applications and business services from the physical implementation of data integration, enabling greater flexibility and optimization of data placement, access and quality. Model driven technologies for SOA and data centric applications are developed separately, modeling the SOA application using Unified Modeling Language (UML) and Business Process Model (BPM) with BPEL, and modeling the databases using conceptual, logical or physical data model. To satisfy the Data Service's requirement, these two areas should be linked together to improve the traceability and agility from Data, IT and business perspectives. In this paper, firstly we improve the traditional data modeling technology and propose a unique Information Liquidity Model (ILM) to take care of different aspects and patterns of data integration. After that, we setup the two-way linkage between our existing SOA service metamodel and new ILM, the true model-driven data service development can be implemented from PIM to PSM.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below