Scalable relational query results handling in service oriented architectures

The focus of Grid computing shifted in recent years towards data-intensive applications. Additionally, it became more standard-based and the adherence to and usage of them has to be further fostered in order to enable and ease exchange and interoperability among the research community. The OGSA-DAI framework for data access and integration together with its default XML output format for relational data, namely Java's WebRowSet, are two important representatives of these developments. The research reported in this paper is being conducted in the context of the EU project ADMIRE. The paper's contribution is twofold and targeted on improving the scalability of query result handling in SoA. First, an XML indexing approach is elaborated to provide an out-of-core Java implementation of the standard Java WebRowSet interface to handle big XML files with low main memory consumption. Second, the concept of fine grained data statistics for data preprocessing tasks calculated on-the-fly data service side is presented.