There was a time when business intelligence was synonymous with intelligence being created long after an activity has taken place. With the business becoming more competitive, there is an inherent need to create intelligence on the fly during the transaction process itself. Business cannot wait for the data to be loaded in the data warehouse during the night batch process and then the analytics routines to run to create intelligence the next day. For this to happen, the BI interfaces need to integrated with the transactional application to provide a seamless feel to the end user.
This seamless integration between transactional and BI systems is brought about by using web services in service oriented architecture. The request for information is passed via a web service to the DW or any other data store and the result which may be analytical reports can be transferred via the same services. This integration of analytical function with transaction is called embedded BI.
Similarly on the data integration side, data can be integrated from multiple data sources in real time. Many companies are using traditional ETL tools in a web service environment to achieve this integration. Many data providers use BI delivery mechanisms wrapped in web services to deliver data to their customers.
For creating these embedded business intelligence applications, understanding both domains of service architecture and business intelligence is important, but the key requirement here is to understand the bigger process in which both transaction and decision making is woven together. The application of this concept is huge and can create really enriched applications.
By vishal.jindal (1) on February 8th, 2010
