A variety of applications can benefit from bidirectional data replication using HDR. Typical applications include:
One of the frequently used applications for data replication is decision support. Typically, these applications do not need the most up-to-date information. For example, it usually does not matter when studying monthly business trends to have data that is up to the second. Also, record-by-record access to data on mainframes must compete with production data processing in these environments. Therefore, it often proves beneficial to transfer a database to a local machine, thereby burdening the production mainframe system with only one access to its database. After transferring a copy of the mainframe data to the SQL Server database in Windows NT, analysts may read and reread the data many times on their local systems without affecting mainframe response time.
Production data processing systems often do not store historical data that may be valuable for trend analysis and similar decision support applications. One solution is to schedule regular HDR replications from DB2 to SQL Server, thereby accumulating needed historical data. Over time, this replication creates a data warehouse that is valuable for historical analysis and other purposes.
Frequently, production systems undergo changes that add and subtract fields in the databases they use. Also, database fields tend to change their definitions over time. Such additions, subtractions, and changes make it difficult to accumulate information over time and make it necessary to clean the data. This cleaning requires some careful massaging of the data in ways that are not easy to accomplish in production systems: making a change, examining the change, testing the resulting database, backing out the change, trying another change, and so forth. Data cleaning is easy and safe to do on data replicated to SQL Server, where there are many tools for constructing and evaluating the process.