Having written a post on topological differences (EAI vs ESB) of integration long time back, and management approaches (centralized vs distributed) I am making an attempt to look at approach in terms of data integration vs application integration.
We have slightly touched upon this subject of Application Integration on one of the earlier post - Binding energy in software systems.
So, what is data integration? A typical Extract-Transform-Load - ETL, data migration, Change Data Capture - CDC, Master Data Management scenarios are classified as data integration. There are several platforms from big and small vendors to achieve this. (IBM InfoSphere/DataStage, Informatica, Microsoft, Oracle, Pervasive have suites/products under this head)
How is this different from traditional Application Integration? Application Integration focuses on integrating business process supporting the information workflows. Data Integration focus is primarily on the propagation and synchronization of data across the enterprise system landscape sometimes spanning into Cloud..
When the volumes are high, a data integration based approach has an advantage; If the process/workflow complexity is high and orchestration is needed to achieve the integration then one should look for an ESB/SOA style integration.
Enterprise Data Integration platforms are comparably similar to EAI platforms in their nature of technical architecture (hub/spoke based EAI topology) with a ETL hub that loads the data into a warehouse.
Of late, the buzz is toward real-time data integration based on CDC etc., let us see how it goes and changes the game!
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment