Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»

The quick-restore capability must be able to restore both full databases and partial databases. The size of the data found in the data warehouse mandates that only partial databases be able to be recovered.

In addition, the DBMS needs to sense that an error has occurred in as automated a manner as possible. Leaving the detection of data corruption to the end user is a very crude way to process. Another useful technology is the ability to create diagnostic tools to determine exactly what data has been corrupted. The diagnostic tool must operate within huge amounts of data.

Other Technological Features

The features discussed here are only the most important. Many others support data warehousing, but they are too numerous to mention here.

It is noteworthy that many other features of DBMS technology found in the classical transaction processing DBMS play only a small role (if they play a role at all) in the support of the data warehouse environment. Some of those features include the following:

■■ Transaction integrity ■■ High-speed buffering ■■ Row/page-level locking ■■ Referential integrity — VIEWs of data

Indeed, whenever a transaction-based DBMS is used in the data warehouse environment, it is desirable to turn off such features, as they interfere with the efficient processing of data inside the data warehouse.

DBMS Types and the Data Warehouse

With the advent of data warehousing and the recognition of DSS as an integral part of the modern information systems infrastructure, a new class of DBMS has arisen. This class can be called a data warehouse-specific database management system. The data warehouse-specific DBMS is optimized for data warehousing and DSS processing.

Скачать в pdf «Building the Data Warehouse»