Data Warehouse Maintenance under Concurrent Schema and Data Updates

Xin Zhang, Elke A. Rundensteiner

Abstract: Summary form only given. Data warehouses (DW) are built by gathering information from several information sources (IS) and integrating it into one repository customized to users' needs. ISs are typically owned by different information providers and hence are independent and autonomous. This implies they will update their data and schemas independently and without any concern for how this may affect the DW defined upon them. Hence, solutions for data warehouse maintenance are needed that can handle such IS updates. We have developed a first comprehensive solution that successfully tackles this problem (X. Zhang et al., 1998). In particular, to guarantee the correctness of concurrent view definition evolution and view extent maintenance of a data warehouse, we introduce a framework called the SDCC (Schema change and Data update Concurrency Control) system. SDCC integrates various algorithms designed to address different individual view maintenance subproblems into one system, such as algorithms for view extent maintenance after IS data updates, for view definition evolution after IS schema changes, and for view extent adaptation after view definition changes.