What is ODS role?

I think that here answer will come from experience.

I开发者_C百科 am right know wondering what is the real goal of an Operational DataStore (ODS) I am browsing the web but if possible I would like to have a real case example feedbacks.

I have few questions :

• ODS is from the Operational side, so what data modeling should be applied to it. Entity-Relationship Modeling but a bit more aggregated or a Dimension Modeling (if so what would be the difference with the data warehouse?) ?

• On most articles it says that ODS posseses real-time or near-real time values, so how often it is feed? On a hourly basis? On a transaction basis?

• What application will be fed by the ODS ? OLAP cube? Reporting Tools? One more time, if those apps goes directly to the ODS what is the data warehouse role?

Any information about general purpose of the ODS and the main differences and the complementarity with the DataWarehouse will be welcome.

Thank you,

An operational data store is the database(s) that the company uses to operate. It's usually modeled as entities and relationships. The logical design of the ODS is normalized, while the actual implementation can be de-normalized for performance. The ODS is designed for quick inserts and updates, with as few indexes as necessary. The ODS contains real-time or near real-time data, but doesn't contain much historical data.

A data warehouse is the database(s) that the company uses to store historical data. It's usually modeled on dimensions. The logical design of the DW is the star pattern, or multiple star patterns. The DW is designed for quick selects, since rows are bulk inserted periodically. The DW contains historical data, and is usually a period or more behind the ODS.