In general, we have 4 Architecture
- ODS – Operations Data source
- NDS – Normal Data source
we can implement this Architecture either in Inmon Model or Kimball Model
Top to bottom Bottom to top
Universally – 8 Architecture
Inmon – Simple, ODS, NDS, Integrity
Kimbell – Simple, ODS, NDS, Integrity.
SRC —- WH
- Inmon Models are preferred when Requirements are very clear.
- Kim bell Model is preferred when Requirements are not clear.
- For the project, we deal with
- ODS –à80% ODS Model
Sol:- when Business Requirements are defined, click/customer requests the team what we want exactly
Project Architecture depends on no. of operational and Analysis Reports.
- More priority is given to operational and Analysis Reports
- Are we have more operational Reports we use ODS.
- Less operational Reports and more Analytical Reports, we use NDS.
- In NDS, we can’t See ODS, But in ODS, we can find warehouse and Data marts (NDS).
- ODS gets refreshed every 10 seconds.
- ODS dim Table will always be type 1 (No History)
Pull You get (take) the data
Push he sends the data
DRA:- Data Repository Area
CSA:- Common Staging Ares
DRA:- Will get only files – Memory which is accessed by different departments.
Data Query:- When the data is being such what the data should contain
Quality Gates:- Requirements .
Every file should have a specific file name:-
To validate the files according to Quality Gates.
Loading, Stage, and warehouse is a Single Data stage.
LoadingStaging ODS Warehousing Box size of 100 Tear loading
10 TB loading
30 TB Staging
60 TB Warehouse.
Dumping the data from DRA to loading à a copy of data
When we load Data to Target data, We maintain all details, that is sequence no;