AUP: Adaptive Change Propagation Across Data Model Boundaries

Although databases focus on the longevity of data, rarely is this data or its structure static. This is particularly true in some domains such as the protein databases that have seen and continue to see an exponential growth rate. Managing the effects of change on derived information (views, web pages) and on applications has been recognized as an important problem. Previous research efforts have developed techniques to both propagate sources changes to views as well as techniques to hide the change from the views and other dependent applications. While this continues to be an active area of research, the problem of management of the effects of change is further compounded by the semantic and the structural heterogeneity that in practice often exists between the source and the derived target information. In this work we now examine and address the problem of change propagation across these semantic and structural heterogeneity barriers. This work is based on our previous work Sangam, which provides explicit modeling of the mapping of one data model to another in the middle-layer. In this work, we now present an adaptive propagation algorithm that can incrementally propagate both schema and data changes from the source to the target in a data model independent manner using the Sangam framework as enabling technology. We present a case study of the maintenance of relational views of XML sources to illustrate our approach.