We’re excited to announce the Public Preview of Cross-Platform View Sharing. Out there as we speak, it permits knowledge suppliers to share views throughout completely different platforms, clouds, and areas, selling an open and interoperable knowledge ecosystem.
View sharing has been helpful; different distributors do it as properly. However till now, it’s largely been restricted to the identical platform. You can share views inside one platform however not throughout a number of platforms and clouds. Databricks solves this drawback with cross-platform view sharing and allows you to share views seamlessly throughout completely different environments. This can be a recreation changer as a result of it expands knowledge suppliers’ attain and avoids vendor lock-in for knowledge shoppers, making collaboration simpler and sooner.
Cross-platform sharing aligns with Databricks’ open sharing imaginative and prescient by enabling safe and seamless knowledge change throughout completely different platforms, clouds, and areas
Understanding View Sharing
To know view sharing, let’s first perceive views. In Databricks, views are read-only representations of information created from tables or different views. They retailer question textual content however not the info itself. Views are a part of the Unity Catalog
View sharing permits customers to share views utilizing the Delta Sharing protocol. Delta Sharing is the trade’s first open protocol for safe knowledge sharing, making it easy to share knowledge with different organizations no matter which knowledge platforms they use. View sharing promotes reusability and reduces redundancy, as a number of customers can entry and make the most of the identical views for evaluation.
Beforehand, when a view was shared between Databricks accounts, shoppers might question it utilizing solely Databricks Serverless SQL. Databricks Serverless SQL works throughout all three main cloud suppliers: AWS, Azure, and Google Cloud Platform (GCP), so views might be shared throughout clouds.
Now, with cross-platform view sharing, knowledge shoppers can leverage any sort of Databricks cluster and even make the most of open Delta Sharing shoppers to entry and question shared views. Open Delta Sharing shoppers are instruments or platforms that assist the Delta Sharing protocol, permitting customers to entry shared views with no need to make use of Databricks. These shoppers embrace in style methods like Apache Spark™, Pandas, Energy BI, Tableau, and others. This makes it potential for customers throughout platforms i.e., who will not be on Databricks, to nonetheless entry and question the shared views by way of Delta Sharing.
Let’s check out this demo to see Cross-Platform View Sharing in motion
Use Circumstances
Databricks to Databricks (D2D) Sharing
On this state of affairs, two Databricks prospects can share views seamlessly inside the Databricks ecosystem. Why is that this necessary? Organizations collaborate with companions who could also be on completely different clouds and in several areas and need to share views with shoppers/companions throughout clouds and areas. By leveraging Delta Sharing expertise, they will seamlessly and securely share views, with out making duplicate copies of the info.
Databricks to Open (D2O) Sharing
On this state of affairs, Databricks prospects can share views with exterior recipients who will not be utilizing Databricks. Cross-platform view sharing helps open connectors (akin to Apache Spark™, Pandas, Energy BI, Tableau, and many others.), permitting recipients to entry shared views by way of the Delta Sharing protocol. This functionality is especially helpful for enterprise analysts and line of Enterprise Customers who require simplified entry to knowledge with no need to work together straight with advanced knowledge platforms
Databricks Market knowledge suppliers profit from cross-platform view sharing by considerably increasing their market attain and monetization alternatives. This functionality permits them to share views with a wider viewers, together with shoppers not utilizing Databricks, thereby rising their potential buyer base. Knowledge Customers will not be restricted to querying views from the Databricks Platform, avoiding lockin with Databricks.
Cross-Platform View Sharing is a game-changer for our prospects. Bringing zero-copy knowledge sharing to advanced enterprises at scale requires flexibility. The power to share views throughout platforms allows us to offer the safety and efficiency advantages of Delta Sharing to extra prospects, serving to them unlock worth from their buyer knowledge sooner
— Derek Slager, CTO and Co-founder of Amperity
What’s forward
Within the coming months, readers can count on Databricks to introduce a number of superior data-sharing options. The upcoming options embrace Sharing for Lakehouse Federation, which permits knowledge suppliers to share knowledge straight from varied platforms (e.g., Amazon Redshift, Azure Synapse, Google BigQuery, Snowflake) with out the necessity for replication.
Moreover, D2O OAuth Assist will improve safety by enabling recipients to authenticate utilizing OAuth tokens from their trusted Identification Suppliers (IdPs). Moreover, the sharing of materialized views and Delta Stay Tables will permit for environment friendly distribution of pre-computed question outcomes and streaming knowledge, offering contemporary knowledge with higher efficiency and decrease prices.
Getting Began
Cross Platform View Sharing is out there in Public Preview as we speak to AWS, GCP and Azure prospects. Study how you should utilize the Delta Sharing open-sharing protocol to share knowledge out of your Unity Catalog-enabled Databricks workspace with any person on any computing platform, anyplace