Choosing between a Fabric Lakehouse or Warehouse
Microsoft Fabric gives teams the flexibility to choose between lakehouses or warehouses or use both. Because both a lakehouse and warehouse have similarities, this can cause confusion when understanding which resource to use.
Similarities
Similarities that exist between lakehouses and warehouses are:
- Both are highly performant and scalable data storage solutions
- Both store data in OneLake in open Delta format
- Both have a SQL analytics endpoint for querying and analyzing data with T-SQL in Fabric or third-party tool
Differences
Now let's get into some of the differences.
Decision Tree
Here are some other considerations when choosing between a lakehouse or warehouse.
Architecture Considerations
Consider the following data architecture scenarios when using the medallion architecture for the various data layers.
Scenario 1: Lakehouse Only
In a lakehouse only scenario, you may use one or more lakehouses to implement your bronze, silver, and gold data layers. You may use Fabric notebooks for your data transformations and data pipelines to orchestrate the data extraction and transformation activities.
Scenario 2: Warehouse Only
In a warehouse only scenario, you may use one or more warehouses to implement your bronze, silver, and gold data layers. You may use T-SQL stored procedures for your data transformations and data pipelines to orchestrate the data extraction and transformation activities.
Scenario 3: Lakehouse + Warehouse
In a hybrid scenario, you may combine the use of lakehouses and warehouses to implement your bronze, silver, and gold data layers. Your bronze and silver layers may use lakehouses combined with notebooks for your data transformations, while your gold layer would use a warehouse with T-SQL stored procedures for your data transformations.
Conclusion
There are a number of factors to consider when choosing between a Fabric lakehouse and warehouse in your data architecture. In general, using a lakehouse will give you the most flexibility with data sources that are both unstructured and structured. Lakehouses also tend to be tailored for experienced development teams that desire a pro code environment. Warehouses provide teams with a strong T-SQL background an option to continue that experience in Microsoft Fabric. You DO NOT need to choose only lakehouses or warehouses when using Microsoft Fabric. Often times both lakehouses and warehouses can be used together in many data architecture scenarios. I hope going through some of these scenarios provided some clarity when choosing between a lakehouse or warehouse in Microsoft Fabric. In my next blog, I will discuss choosing between data pipelines and dataflows in Microsoft Fabric. Thanks for reading and happy continued learning!
© Copyright. All rights reserved.
We need your consent to load the translations
We use a third-party service to translate the website content that may collect data about your activity. Please review the details in the privacy policy and accept the service to view the translations.