Trigger Real-time Business Processes with Adabas Change Data Capture

Issue 3, 2012    


Getting information to the right person, at the right time, is a core objective of IT organizations.  Event Replicator for Adabas on LUW (Linux, UNIX, Windows) facilitates real-time business processes, data warehouses, data analytics, and disaster recovery for Adabas

Companies need to increasingly drive business processes based on real-time changes to data. This is a challenge for large organizations where information is spread across multiple operational data stores. Most companies have a minimum of 10 different database types as a result of acquisition(s) or using off-the-shelf products tied to a specific database. Gleaning meaningful information from these disparate data sources requires combining the data.  In the past, this usually meant building a monolithic data warehouse that was difficult to keep up-to-date due to the volume of data. Populating a data warehouse takes a lot of time, in some cases 10 to 24 hours. Refreshing the data in the warehouse typically occurs during a nightly batch window, making the warehouse unavailable during that time. In addition, making confident daily or intra-day decisions based on the data warehouse information that is typically a day old or more is challenging or near impossible.

Change data capture (CDC) provides a mechanism to drive business processes in real-time, based on changes to production operational data stores. Event Replicator for Adabas on LUW, using CONNX technology, provides real-time change data capture for Adabas and solves the business problems caused by disparate data sources. With Event Replicator, you can:
  1. Keep Adabas data in sync with a relational database for real-time analysis.
  2. Push changed Adabas data to JMS and the webMethods Integration Server to drive business processes in real-time.
  3. Keep a disaster recovery nucleus for Adabas in sync with the production nucleus.


FIGURE 1:  Event Replicator provides real-time change data capture
 

Real-time Data Analysis  
There are many tools on the market for performing deep data analysis and data mining. These tools typically require data intensive queries. By pointing these data analysis tools to a “live relational clone” of the data, impact to the production Adabas nucleus is negligible. Data can be pushed to a data mart or data warehouse in real-time, eliminating the need for a costly nightly batch extract and upload. The same source of Adabas data also replicates to multiple target databases simultaneously as needed. 

 
Trigger Business Processes in Real-time
For data stored in a relational database, a SQL “trigger” provides a mechanism to perform business logic based on a change to a record in real-time. Until recently, if the data was in Adabas on LUW, a user exit was the only way to provide change data capture (CDC). User exits, typically written in C or assembler, have an impedance mismatch with languages such as Java, Perl, or Python making it more difficult to utilize these languages in a “trigger” process.  Now, there is a new solution.  Event Replicator captures changes and pushes them in real-time to a Java Message Service (JMS) Queue. This is ideally suited for pushing changes to the webMethods Integration Server for further processing of business logic on the Enterprise Service Bus (ESB).
 
 
Disaster Recovery
Event Replicator provides a new method of maintaining a cloned Adabas nucleus that is synchronized with the original.  This real-time, sub-second synchronization enables a hot backup, or disaster recovery nucleus to remain up to date at all times. If there is a failure event on the machine that contains the primary nucleus, applications can switch to the backup nucleus and continue work. Once the failure on the primary system has been resolved, the data from the disaster recovery nucleus can be restored to the primary nucleus, and real-time synchronization continues.
 

A Versatile Tool
The three scenarios above are not mutually exclusive.  Event Replicator can be used to replicate to a relational database, push changes to a JMS queue, and keep a disaster recovery nucleus synchronized: simultaneously or in any combination.   This makes Event Replicator, with CONNX, the most versatile and powerful tool for propagating Adabas data. The replication technology is built directly into the Adabas nucleus – so there is no external scanning of PLOGs, no user exits.  This provides a tightly-coupled, highly-reliable replication mechanism.  
 
Learn more about this versatile tool by visiting the Tech Community and reading the Wiki tutorial on Event Replicator for Adabas on LUW