Derived data for decision support systems

摘要

In decision-making situations requiring “what-if” analysis, in statistical databases, and in distributed databases, it is desirable to explicitly store derived data without losing consistency with the original data. We introduce independently-updated views (IUVs) for storage and use of derived data for decision-support systems. IUVs support multiple versions, provide derivation transparency, maintain data consistency, and afford control over time of derivation. The notion of data consistency is extended to allow for multiple versions so that versions are consistent with the actively updated database on which they are defined. Implementation issues are discussed and the cost of retrieving IUVs representing different types of derived data is estimated.