Retrieval of Heterogeneous Data from Dynamic and Anonymous Sources

Distributed knowledge is both a valuable asset and difficult to utilize. Uncountable data sources in many different fields (e.g. science, business, education) are used mostly isolated. Trying to combine and mine their wealth of information often rises the problem of missing a central link to mediate between differently designed data systems. In dynamic environments with data sources, that can change their availability and structure over time, the management of a central mediator-based system is too static. We propose an information system that transfers the management of data sources to the client-side, allowing data owners to keep full control over the third-party data access while providing simple administration. Data owners can register their data sources directly. Queries are forwarded to be translated on the data source level. In contrast to a centralized mediator-based system, which is not suitable for dynamic retrieval and flexible adjustment of heterogeneous data, our system provides the adaptability to keep query responses accurate in case of evolving data sources.