Managing Virtual Data Marts with Metapointer Tables

Although Internet/Intranet topology has radically evolved communications for information workers, the access and transfer of large volumes of data through this topology is still inefficient and is not inherently conducive to point-in-time recall. Because information managers make real time decisions based on relevant subsets of an enterprise data warehouse, data marts are required for optimized data accessibility and use. Storage virtualization, based on pointer systems, has introduced new instant copy functions which do not require that data be actually stored outside the data warehouse even though it is referenced in a variety of contexts (e.g., multiple data marts utilizing overlapping subsets of the data warehouse). The purpose of this paper is to describe a method enabling users to access relevant subsets of large data populations residing in different data warehouses. This method enables the creation of data marts for use by local decision support software without requiring the data mart data population to be stored locally. This will save storage space and data transmission resources, and simplify synchronization of the data mart and the warehouse when updates occur.