Comparison between capacity units (CU) in Fabric and data warehouse units (DWU) in Azure

[Ostrzak]

Hi,

I would need a rough estimate of computing power comparison between Fabric and Azure. 

Currently Fabric has rolled out new pricing system, where subscription plans are measured in capacity units (e.g. F256 offers 256 CU) available. On the other hand, in Azure dedicated SQL pool also offers a fixed amount of computing power, conceptualized as data warehouse units ( e.g. DW 1000c offers 1000 DWU).

I cannot find a way to compare both, even roughly. I understand that these are umbrella-concepts for collection of different analytic resources, but there still should be a meaningful way to compare them.

Thank you in advace for your help.

[v-nikhilan-msft]

Hi @Ostrzak 
Thanks for using Fabric Community.

Fabric:

• Subscription-based: Offers multiple tiers (e.g., F256, F512) with varying capacity units (CUs) representing available resources.
• Pay-as-you-go model: Resources are dynamically allocated based on usage within your subscribed capacity.
• Bundled services: Includes access to various data and analytics services like Synapse Analytics, Data Factory, Databricks, and others.

Synapse:

• Multi-faceted pricing: Combines pay-per-query for serverless SQL pools, hourly charges for dedicated SQL pools, and additional costs for data storage, data transfer, and managed services.
• Flexible options: Offers various dedicated SQL pool tiers (DW100c, DW2000c, etc.) with fixed resources and serverless pools with scaling based on query demands.
• Stand-alone service: Primarily focused on data warehousing and analytics but integrates with other Azure services.

Approaches to a Estimation:

• Focus on Specific Workloads: Identify your specific workload and compare performance benchmarks for similar tasks between Fabric and Azure offerings (e.g., running queries on large datasets).
• Consider Resource Composition: While not directly comparable, analyze the advertised resource composition of CUs and DWUs for similar tiers (e.g., compare CU composition for F256 with DWU composition for DW1000c). This might offer approximate insights.
  

Please refer to these links for more information:
https://endjin.com/blog/2023/05/azure-synapse-analytics-versus-microsoft-fabric-a-side-by-side-compa...

https://www.pythian.com/blog/understanding-the-differences-between-azure-synapse-analytics-and-micro...

https://blog.fabric.microsoft.com/en-us/blog/microsoft-fabric-explained-for-existing-synapse-users/

Hope this helps. Please let me know if you have any further questions.

[v-nikhilan-msft]

Hi @Ostrzak 
We haven’t heard from you on the last response and was just checking back to see if you have got some insights regarding the query. Otherwise, will respond back with the more details and we will try to help.
Thanks

[v-nikhilan-msft]

Hi @Ostrzak 
We haven’t heard from you on the last response and was just checking back to see if you have got some insights regarding the query. Otherwise, will respond back with the more details and we will try to help.
Thanks

[Ostrzak]

Hi V. Nikhilan Thank you for your first answer. The linked articles explain well the difference between Fabric and Azure, but don't answer my question regarding computational power comparison. I don't have a possibility to run comparable workload in both services, nor time to read through documentation regarding resource composition of CU and DWU.

If you could provide any additional insight, I would be grateul.
Thanks

[v-nikhilan-msft]

Hi @Ostrzak 
Apologies for the delay in response.

I will request you to please go ahead with Microsoft support on this support.
After creating a Support ticket please provide the ticket number as it would help us to track for more information.

Hope this helps. Please let us know if you have any other queries.