Operational databases are designed to follow processes dictated by their objectives according to business rules. In essence this means that operation databases follow prescribed procedures and produce and store data that fulfill the operational business objective. Operational data changes constantly (and therefore does not exist in a single state for a long time), is low in granularity and has a single focus (process and relationships). As a result operational data is difficult to understand as a whole without interpretation (due to its tabular nature) however it is possible to estimate the transactional value, storage and technical requirements of an operational database.
Decision support databases, on the other hand, use the data from the operational database with external and business model data to allow the end user to perform queries that form as basis for decision making. The data for and results of these queries can be stored in the decision support database. As a result the data within a decision support database reflects the nature of the data required to make decisions, it covers a period of time that ties in with decision making timeframes and has a higher level of granularity which is easier to understand and more flexible in presentation (allowing reports to be built across operational database tables over time, for example).
Due to the differing nature of operational and decision support databases in actions, transactional and query terms, different data stores usually make objective sense. The data with operational databases changes frequently and transaction are much higher when compared to that in decision support databases and results in highly normalised relational tables (Coronel (2009)). Whereas the queries performed on decision support databases tend to be bespoke and/or much more complex than with operational databases and a result a less normalised data approach makes queries less complex (e.g. star schema).
With these differences then a separate data store for the operational database and the decision support database is a natural choice to take account of data volatility, subject orientation, time factor and integration differences. If the system used the same data store then the conflicting nature would cause issues that would affect the likely overall success of each database.
Coronel, Morris & Rob (2009) Database Systems: Design, Implementation, and Management (9th Edition). Cengage Learning.
Information Management Magazine (1998) The Operational Data Store [Online]. Available at http://www.information-management.com/issues/19980701/469-1.html (Accessed 23 May 2010).
Power, D (2007) How does DSS data differ from operational data? [Online]. Available at http://dssresources.com/faq/index.php?action=artikel&id=146 (Accessed 23 May 2010).