数据仓库概念.pptVIP

38 2 1 38 2 1 38 2 1 Designing a successful data warehouse solution poses a number of challenges. The design must produce a system which is : Extensible. It must be possible to add new types of transactional data and introduce new aggregations and summaries to a warehouse as information requirements evolve through usage. Scaleable Data warehouses are capable of storing and managing enormous volumes of transactional and summarised data. A data warehouse may grow by an order of magnitude over time through the need to add new types of data or because historical data must be retained on-line for long periods. Flexible The data access paths to detailed and summarised data must be flexible to support multi-dimensional analysis, drill down, and true ad hoc querying. It must be possible to explore the data in a way which enables theories to be developed and tested. Integrated A data warehouse must be fully integrated with the existing systems and operational environment so that data from multiple sources can be loaded efficiently Reliable All discrepancies in data loaded from multiple sources must be reconciled so that the warehouse always contains information that is accurate and consistent for a given point in time. In a recent survey undertaken with over 4,000 organizations, the overwhelming results were : Deciding on the most appropriate tool for the job, is one of the most difficult decisions made when implementing a Decision Support System. The data access tool is what makes or breaks the success of the decision support system project because it is most visible to the people who benefit from that decision support system. Of the organisations surveyed, several key criteria emerged for selection of a decison support development tool: Note that Ease of Use is by far the most important criteria. Performance and Analysis could also be classified as Usability issues. Centralized administration is key for low maintenance, setup, control and scalability. Established Ve