Learn about Databricks Pricing across different cloud providers – AWS, Azure , GCP

Databricks Pricing (Cost) depends on the cloud provider of your choice – AWS, Azure , GCP

Databricks Cluster Pricing (Cost) consists of two components – one will be Databricks Platform costs and other will be Compute resources associated to your Cloud Provider.

Databricks Cluster pricing (Cost) is dependent on following combinations

1. Subscription plans
   1. Standard – Platform for Data Analytics and ML Workloads
   2. Premium – Platform for Data Analytics and ML Workloads at Scale.
   3. Enterprise – Platform for Data Analytics and ML Workloads for mission critical applications. Majority of the customers will fall under Enterprise.
2. Compute type
   1. Jobs Light Compute – Equivalent to Opensource Spark. Ideal for non-critical workloads.
   2. Jobs Compute – Data Engineering pipelines to build and manage data lakes.
   3. SQL Compute – SQL Queries for BI Reporting, analytics, Visualization on data lakes.
   4. All-Purpose Compute – Data science, Machine Learning, Data Engineering, BI and Data analytics workloads.

Databricks Unit – DBU

Databricks Unit (DBU) is a unit of processing capability per hour, billed on a per-second usage.

DBU/hour is driven by the Databricks Runtime version and the driver & worker sizing.

Databricks recommends “i” series (AWS), “l” series (Azure), “n2” series (GCP) instance types as they provide accelerated data access through Delta caching.

AWS Databricks Pricing

Following is sample Databricks Pricing (Cost) calculator if you utilize a Databricks cluster of instance type family “i3”. Estimator assumes a cluster with 1 Driver node and 3 Worker nodes . It also assumes “Enterprise” plan with “All-Purpose” Compute Option.

Following is the summary

More details to be found in – https://databricks.com/product/aws-pricing

Azure Databricks Pricing

Azure has only Standard and Premium plans.

Following is sample Databricks Pricing (Cost) calculator calculator if you utilize a Databricks cluster of instance type family “L”. Estimator assumes a cluster with 1 Driver node and 3 Worker nodes . It also assumes “Premium” plan with “All-Purpose” Compute Option.

Following is the summary

More details to be found in – https://databricks.com/product/azure-pricing

GCP Databricks Pricing

GCP has only Standard and Premium plans.

Following is sample Databricks Pricing (Cost) calculator if you utilize a Databricks cluster of instance type family “n2”. Estimator assumes a cluster with 1 Driver node and 3 Worker nodes . It also assumes “Premium” plan with “All-Purpose” Compute Option.

Following is the summary

More details to be found in – https://databricks.com/product/gcp-pricing

Conclusion

In all the above calculations, instance types across multiple cloud providers are in the same CPU/Memory range.

As shown above, Data bricks Platform cost is high given its rich features, functionality, performance and ease of use.

Among all the Cloud service providers, AWS compute costs is on the lower end with Azure on the high end.