Consistency Algorithms for Multi-Source Warehouse View Maintenance

  • Authors:

  • Yue Zhuge;Hector Garcia-Molina;Janet L. Wiener

  • Affiliations:

  • Computer Science Department, Stanford University, Stanford, CA 94305;Computer Science Department, Stanford University, Stanford, CA 94305;Computer Science Department, Stanford University, Stanford, CA 94305

  • Venue:
  • Distributed and Parallel Databases - Special issue on parallel and distributed information systems
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

A warehouse is a data repository containing integrated informationfor efficient querying and analysis. Maintaining the consistency ofwarehouse data is challenging, especially if the data sources areautonomous and views of the data at the warehouse span multiplesources. Transactions containing multiple updates at one or moresources, e.g., batch updates, complicate the consistency problem. Inthis paper we identify and discuss three fundamental transactionprocessing scenarios for data warehousing. We define four levels ofconsistency for warehouse data and present a new family ofalgorithms, the Strobe family, that maintain consistency as thewarehouse is updated, under the various warehousing scenarios. Allof the algorithms are incremental and can handle a continuous andoverlapping stream of updates from the sources. Our implementationshows that the algorithms are practical and realistic choices for awide variety of update scenarios.