Principles of Data Integration

10

Data Warehousing and Caching

Thus far, we have discussed the key approaches to defining schema mappings, composing queries with schema mappings (reformulation) to get queries posed directly over data sources, and query processing over remote data. These basic techniques enable us to fetch data on demand in response to user information needs, using the most up-to-date state available. However, in a variety of circumstances this fully virtual data integration architecture may not be ideal: we may be willing to tolerate slightly older data in exchange for better performance, better data quality, or the ability to express more complex queries or perform more sophisticated data transformations. The basic idea is to exploit materialization — i.e., ...