Applications generally require sending data from the mainframe to the distributed systems and also consolidating data from the distributed systems to the mainframe. HDR satisfies both requirements. With HDR, you can move data both ways between DB2 and SQL Server databases.
HDR operates on whole tables in a relational database. It performs "snapshot" replication — that is, it copies all or part of a database table to a destination table. It does not perform incremental replication because it has no log reader component and does not attempt to track individual updates to the table. Although it does not support incremental replication, you can often get most of the desired effect with HDR if the source data has a sequence number or timestamp to identify the most recent rows in the table.
Although SQL Server provides a "publish" or "push" replication capability, it is intended for use among SQL Servers only. HDR adds the ability to "pull" relational data from the mainframe database. HDR can access data stored in DB2 relational databases on the full range of enterprise computing platforms: mainframes, midrange machines, minicomputers, and distributed workstation servers. However, HDR cannot access other kinds of mainframe files such as IMS, VSAM, or partitioned data sets. Likewise, HDR can access relational data in SQL Server databases, but not other kinds of Windows NT Server files.
When you configure a replication operation, HDR queries the source database and automatically communicates data type and attribute information to the target database along with the actual data. You can also configure custom conversions for exceptional cases.