You can use different Host Data Replicator features and configurations to provide the data replication that best supports your enterprise application.
HDR supports replication of selected columns (vertical filtering), selected rows (horizontal filtering), and replication of selected columns from selected rows (combined vertical and horizontal filtering). The replicated data can either replace existing tables or be merged on a row-by-row basis. HDR also supports derived destination columns and data transformations, including automatic conversion of date and timestamps.
You can use HDR to subset the database being replicated. Subsetting may be horizontal (rows), vertical (columns), or both. For example, an analyst may want to look at sales data for only July and August, and would select only the rows of the database corresponding to these months. Similarly, the analyst may want to look only at the sales of a particular product, and would select only the columns of the database corresponding to that product.
Once selected, the database subset in SQL Server is easier to manage and use than the large database superset in DB2. Most analysts running decision support applications understand and appreciate the value of subsetting and use the technique frequently. However, the technique is also valid for production data processing applications. For example, a desktop sales application might use HDR to download a subset of the company's price list every day. The application might require prices of only certain products (vertical subsetting) and only current pricing (horizontal subsetting).
HDR controls replication when you want it. Whether it's a one-time replication or a recurring event at predefined times, you can schedule replication to meet the requirements of your application.
With data replication, a Windows-based application can:
Notice that data replication does not support the data locking and two-phase commit operations that are necessary to ensure correct synchronization of data stored in Windows NT Server and DB2. However, for many casual data processing requirements, using periodic two-way replication is a convenient way of managing distributed data.
For More Information If your application requires two-phase commit transactions, you should consider using the SNA Server Component Object Model Transaction Integrator (COMTI). For more information on using COMTI, see Integrating with Mainframe Transactions in this chapter.
A Windows-based application usually has two ways to access DB2 data. One way is to replicate the data from DB2 to SQL Server and then operate on the data in the SQL Server environment. HDR supports this approach. A second way is to issue SQL commands to the remote database. The ODBC Driver for DB2 supports this approach. A client application issues ODBC commands to the ODBC Driver for DB2 to send SQL statements to DB2. The ODBC Driver for DB2 uses the SNA Server application programming interface (API) to communicate to DB2 through LU 6.2 sessions. This direct connection configuration complements the HDR configuration. For example, an application such as Microsoft® Excel or Microsoft® Access might need certain information in real time or may need to update the mainframe information immediately. However, the same application might use replicated data in the SQL Server for certain information that is not as time-sensitive.