Real-time Database Replication
HVR supports asynchronous log-based capture out of Oracle, SQL Server, IBM DB2 and Ingres. On a well-configured system HVR captures at the tail end of the online transaction log capturing transactions as soon as they are committed to the log. If for whatever reason something fails in the change data capture then HVR will simply restart and resume where it left off with no loss of transactions.
At the destination HVR uses optimizations to ensure scalability for large data volumes irrespective of the type of target database:
- Combining transactions: multiple small transactions become fewer large transactions.
- Coalescing changes: multiple subsequent changes to the same row are combined into a single row change.
- Burst mode: changes are staged, coalesced and then frequently applied using bulk SQL statements to the actual target table(s) modifying multiple rows at the time.
HVR delivers flexible data replication in heterogeneous environments, with support for:
- Automatic mapping between source and destination data types, including automatic mapping of character sets and code pages.
- Bi-directional or multi-directional replication, including conflict detection and resolution.
- Transformations including table and column name changes and derived values at column level.
- Built-in transformations to support soft deletes last updated columns for slowly changing dimensions.
- Initial load capabilities on live systems to ensure systems are in sync once real-time replication catches up to current.
- Compare functionality include detailed reporting and SQL generation in case systems are out-of-sync.