Enterprise Information Management

Now let’s put some thought around the following realities.

1.       BI solutions traditionally are being used by mostly tactical and strategic level users to analyze and monitor various aspects of business health and take off-line decisions. This creates a time lag between the information is viewed; decision is taken and put back to use in the business processes. Therefore, BI solutions are always a downstream system with one way traffic of flow of information.

2.       BI aims at creating a single version of truth of Organization’s information assets and used by business to run their not only for strategic and tactical business decisions but also operational decision making process which demands quick (in some cases instant) decisions.

3.       There are 2 distinct part of a BI solution, namely, Information acquisition and Information exploitation. In both cases, sources and destinations may be unknown. For example, Pharma companies need to buy sales data from external data service providers and also have to share data to outside providers for crunching due to regulatory reasons.

Scenarios such as above make it imperative to have BI solutions loosely coupled from the following standpoint.

1.       BI solution should be able to acquire information and consolidate it in a loosely coupled manner, meaning that it should be able to acquire data from multiple divergent sources and should be able to deliver similar divergent targets. Also, such processes should be event driven so that information is obtained in real time. ETL tools are moving in this direction whereby source and/or target can be a webservice.

2.       Similarly BI solution should be available as a service for the downstream Information consumer. For instance, a call center executive should be able distinguish between a profitable customer and an average customer when he/she receives a complaint so that appropriate prioritization can be done. Such situation is possible if BI can be exposed as a service rather than duplicating the same information at multiple places. BI reporting tools are moving in this direction.

Hence, I believe we will see increased affinity between SOA and BI architectural paradigm which are complementing each other making a closed loop between transactional and BI systems. Do you still see a conflict between BI and SOA?