An operational data store usually stores and processes data in real time. An ODS is connected to multiple data sources and pulls data into a central location.
The way operational data stores work is comparable to the extract, transform and load (ETL) process. ODS systems import raw data from production systems and store it in its original form.
In the ETL process, data is extracted from target sources, transformed and loaded to its destination. In the ODS process, data is not transformed, but rather it's presented as is to business intelligence (BI) applications for analysis and operational decision-making.
In some cases, data from an ODS is replicated and then ETL is used to transport the replicated data to a data warehouse.
As operational data stores ingest data, new incoming data overwrites existing data.
Source : Techtarget