Layered Architecture of EDW
1) DATA MART LAYER
It is a reporting layer to support the known.
It can be used for tactical or strategic reporting with highly aggregated data.
It can be used for operational reporting where data is less aggregated.
Unlike Operational DataStore, the data is integrated because it is loaded form EDW layer.
Data Schemas – Multiproviders, Infocubes
2) ENTERPRISE DATA WAREHOUSE LAYER
It is a corporate memory.
· Single point of truth for auditing
Allows Vertical consistency i.e. the data flow of each master record and transaction record in data mart is traceable.
· Completeness
Date is granular and not aggregated.
Data is not overwritten but new information can be added.
Provides complete history
· Controlled Extraction and data staging
Extracted only once and deployed to several data marts.
· Integration
Data is integrated.
Data Schemas – Data Store Objects
3) OPERATIONAL DATA STORE Layer
The "Operational Data Store Layer" typically comprises operational data that is closely
related to the transactions.
· Fast loading of mass data with a relatively short lifespan, but a high load frequency.
· List Reporting at document level.
· Data distribution to subsequent data targets
Data Schemas – Data Store Objects
