e Traditional Data Warehouse Attributes As the data warehouse technology becomes a mainstream technology, some traditional attributes are being deviated from in order to meet users' increasing demands. The most noticeable ones are timing variant, nonvolatile, and built from scratch. Deviation from time variant & nonvolatile As the size of the data warehouses becomes larger and larger (e.g., in terabytes), the amount of time to reload or append data can become very tedious and time consuming. Furthermore, users are demanding more up-to-date data to be included in the data warehouse. Instead of adhering to the traditional data warehouse attributes of keeping the data nonvolatile and time variant, new data is being added to the data warehouse on a daily basis, if not on a real-time basis. Thus, new approaches are being made to tackle this task. Two possible methods are: Perform hourly/daily batch updates from shadowed log files. Transformation rules are executed in this process. Thus, when the data reaches the target data warehouse database, it is already transformed and summarized. Perform real-time updates from shadowed log files. Again, transformation rules are executed in this process. Instead of batch updates, this takes place on a per transaction basis that meets certain business selection criteria.Deviation from built from scratch For customers that are in the "horizontal" industry, meaning their applications are unique to their own businesses, it is essential to build a data warehouse from scratch. However, for customers that are in a "vertical" industry, meaning their applications are either coming from the same vendor or the functionality of those applications from various vendors are similar in nature, it is possible to leverage an "off-the-shelf" pre-packaged MART. The MART is a data-modeling template that is designed with a certain set of queries in mind for that specific vertical industry. Instead of designing data models fro...
