Also your production database is optimized for different workloads than your analytics database.
Usually production is used for fetching and updating a small number of records at a time (think updating a shopping cart), and has strict latency requirements whereas analytics involves reading a large amount of data in columns (think count group by one or two columns), and can be done in batches where the results can get a more and more stale until the next batch runs.
Comments
Also your production database is optimized for different workloads than your analytics database.
Usually production is used for fetching and updating a small number of records at a time (think updating a shopping cart), and has strict latency requirements whereas analytics involves reading a large amount of data in columns (think count group by one or two columns), and can be done in batches where the results can get a more and more stale until the next batch runs.
How do you batch write the results (say updating shopping carts) when frontend has to reflect whats in the database?
They're talking about moving data between two different back end databases. Your production database optimized for your application/latency.
Then you have your warehouse database that you updated once a day with information from prod.