Querying Heterogeneous Information Sources Using Source Descriptions

摘要: We witness a rapid increase in the number of structured information sources that are available online, especially on WWW. These include commercial databases product information, stock market real estate, automobiles, and entertainment. would like to use data stored these answer complex queries go beyond keyword searches. face following challenges: (1) Several store interrelated data, any query-answering system must understand relationships between their contents. (2) Many not full-featured database systems can only small set over (for example, forms WWW restrict one (3) Since is very large, effective techniques needed prune accessed query. (4) The details interacting with each source vary greatly. describe Information Manifold, an implemented provides uniform access heterogeneous collection more than 100 sources, many them IM tackles above problems by providing mechanism declaratively contents query capabilities sources. There clean separation declarative description actual source. algorithms descriptions effciently for given practical generate executable plans. plans we inolve querying several combining answers. also present experimental studies indicate architecture used Manifold scale up well hundred