Abstract:
Today, many web-services rely on queries over distributed data sources. However, these applications suffer from the unavailability of underlying data sources. In the pres...Show MoreMetadata
Abstract:
Today, many web-services rely on queries over distributed data sources. However, these applications suffer from the unavailability of underlying data sources. In the presence of unavailable data sources, various techniques exist. Worst solutions simply ignore the problem, or generate an error, which is relatively frustrating for the user. Other works consider partial answers or mutant query plans to delay the querying of unavailable sources. Despite of their efficiency, those works are not order sensitive. It means that the data order in the final result does not reflect the underlying data source order and the order defined by the query. In this paper, we propose an order sensitive approach for the evaluation of queries when data sources are unavailable.
Date of Conference: 28-31 October 2007
Date Added to IEEE Xplore: 31 January 2008
ISBN Information: