About the Post

Author Information

RDBMS Replication & Real-Time Tool launched

Database replication has provided an alternative for Enterprises
requiring access to any or all Corporate data without impacting their
normal transaction processing systems. Selected data for ETL, Data
Warehouses or Data Marts is placed and maintained in replicas of the
production databases. These replicas can contain all data or subsets of
the production databases in the same or different RDBMS.

In many organizations, departmental and satellite computing have resulted in the proliferation of different RDBMS implementations. These RDBMS may be the results of personal choice, availability of specific data analysis and reporting tools and RDBMS costs.

Regardless of the RDBMS, it is imperative to provide a fast, efficient, recoverable and proven method of capturing the transaction data from one Oracle database and applying this data to other Oracle or different RDBMS databases on different platforms be they IBM Mainframes, Linux, MS Windows and Unix.

Problem Statements

Ensuring selective, accurate capture and guaranteed delivery of transaction data that has been applied to the source Oracle database with the added requirement to apply the updates across various networks and platforms to the target RDBMS including Disaster Recovery (DR) sites in as near real-time as possible without any manual intervention.

Our Solution

Our application provides a Blackbox rules-based facility to:

- capture the Oracle database transactions from Logs as SQLs,
- selectively extract only the columns required by the replication,
- transform the data (SQLs) where necessary for non-Oracle
targets,
- block the DARSTRAN Oracle Logminer generated SQL in
IBM MQSeries packets
- transmit the packets via an MQSeries Queue

In addition, for users with DR sites,

- DR database(s) can be kept in sync with the production
environment up to and including the last successful COMMIT to MQSeries.
- An immediate Network switch can be done - without waiting
for delivery, loading latest production backups, and running catch-up/resync processes
- All new transactions applied at the DR site can be captured and
used to resync the production database when the production site is back online

And at the target system

- receive the packets
- apply the unpacked SQLs to the target DB

All this in near real-time

Benefits

- Oracle DB (source) changes are captured dynamically from the Logs, real-time formatted (transformed) and transmitted as SQL statements to the target DB through MQ.
- Can be used to provide Disaster Recovery by making the DR site one of the target DBs
- The SQLs are received at the target machine from MQ and real-time applied to target DB.
- Offline capture and replication using archived logs is also possible.
- Failover/Recapture facility in case of errors.
- Detailed logging and reporting facility (including DB monitoring & tuning).
- Dynamic transformations supported for a variety of target databases including IBM DB2 & UDB, CA-IDMS, Informix, MS SQL Server, SYBASE and Oracle
- The source Oracle DB can be any, from Oracle 9i upwards (9i, 10g, 11 etc.)
- Supports different target & source platforms including

IBM mainframes AS400
Unix IBM AIX
Linux HP-UX versions 10 and 11 PA-RISC & Itanium
Windows

Comments are closed.