A data warehouse can never be accessed on a millisecond basis. Because of the nature of the data, the volume of the data, and the mixed workload that uses data from the data warehouse, the data warehouse is not geared to support OLTP types of processes. Guaranteed subsecond response for an access to the data warehouse is architecturally not a viable option.
But subsecond response time is, in fact, very valuable in many operations. Many businesses require very fast response time, and those businesses cannot have access to the data warehouse. When subsecond response time is required and integrated DSS data must be accessed, there is a structure known as an operational data store (ODS) that is the place to go to when high-performance processing must be done.
Unlike the data warehouse, the ODS is optional. Some organizations have an ODS and other organizations do not have an ODS. The need for an ODS depends entirely on the organization and the processing that is being done.
In many ways, the ODS and the data warehouse are complementary structures. Both reside outside of the operational environment. Both support DSS processing. Both use integrated data. For these reasons, the ODS and the data warehouse are often considered to be complementary. Data flows between the ODS and the data warehouse in a bidirectional manner. On some occasions, the ODS feeds the data warehouse. On other occasions, the data warehouse feeds the ODS. ...