An algebra for capability object interoperability of heterogeneous data integration systems

Abstract

In heterogeneous data integration systems, there can be a wide variety of different data sources, which support query interfaces with very varied capabilities. Constrained by the expressiveness of the access interfaces, data integration systems from multiple sources have to cope with different and limited capabilities of the sources. At the same time, several contemporary integration systems export their mediator capabilities in the same way as those of data sources, thus making it hard for them to be integrated by other mediators. In this paper, we propose a capability object conceptual model to capture a rich variety of query-processing capabilities of sources' and outline an algebra to compute the set of mediator-supported queries based on the capability limitations of the sources they integrate. Our work highlights the capability computing approach that enables interoperation among various sources by computing the capability objects associated with them towards scalable application integration. Not only is the process of interoperation semiautomatic and timesaving, but also, often, the end-users have idea of the capabilities of mediator and get a better insight on which queries to submit to the mediator. Finally, we show the properties of the algebraic operators. Query optimization is enabled based on the properties of the algebraic operators. © Springer-Verlag Berlin Heidelberg 2005.