Deliver Bi-Directional Integration for Oracle Autonomous Database and Databricks

Until now, sharing data between enterprise systems often meant complex pipelines, duplication, and lock-in. With Oracle’s support for Delta Sharing, that’s no longer the case. Oracle Autonomous Database—along with Oracle Fusion Data Intelligence—can now securely and seamlessly share data with Databricks and other platforms, all without copying data or breaking governance rules. This blog explores why Oracle adopted Delta Sharing, how it improves collaboration across ecosystems, and the real-world use cases it enables for both technical teams and business users.

Understanding Oracle ADB and Delta Sharing

Oracle Autonomous Database (ADB) is a fully managed, serverless database built for transactional and analytical workloads. It handles provisioning, scaling, and tuning automatically, freeing up teams to focus on building applications and insights, not managing infrastructure. ADB is designed to deliver the performance and reliability enterprises need without the operational overhead.

Delta Sharing, on the other hand, is an open protocol for secure data sharing across clouds, regions, and platforms. Instead of duplicating data or setting up custom APIs, teams can share live data directly, whether it’s with internal teams, external partners, or customers. Built into Databricks Unity Catalog, Delta Sharing ensures that shared data remains secure, governed, and easy to discover for data scientists and analysts alike. It is a fully managed, serverless database that automates provisioning, scaling, and tuning for both transactional and analytic workloads. Designed to simplify data management, it delivers high performance for enterprise applications.

Why did Oracle choose Delta Sharing?

Oracle Autonomous Database customers increasingly need to share data with partners, suppliers, or analytics platforms—quickly, securely, and without creating data silos. For example, a manufacturing company using Oracle ADB to manage product lifecycle data may need to provide real-time visibility to suppliers. Similarly, a retailer storing transactional data in ADB may want to feed data into Databricks for advanced analytics and machine learning.

Historically, these types of data exchanges relied on outdated methods such as FTP, email, or ETL pipelines. While familiar, these approaches often resulted in data duplication, increased storage costs, and delayed insights. More importantly, they were not designed for real-time access, making collaboration cumbersome and inconsistent.

Oracle sought a better path—one that aligned with its commitment to openness and customer choice. Many modern data-sharing solutions functioned as closed ecosystems, locking customers into vendor-specific approaches. Delta Sharing provided a clear alternative: an open, cloud-agnostic protocol designed to break down those barriers.

Here’s why Delta Sharing stood out:

1. Embracing Open, Multi Cloud Collaboration
   Many organizations operate in multicloud environments, using Oracle ADB alongside platforms like Databricks or tools on clouds, such as OCI, Azure or GCP. Delta Sharing enables governed data sharing across these platforms without the complexity of ETL or the overhead of data replication. 
    
2. Support for Business Intelligence Tools
   Business teams rely on tools like Power BI and Tableau to get insights from data. Previously, connecting these tools from Oracle ADB meant building and maintaining custom connectors - a time-consuming process. Delta Sharing simplifies this with built-in support for BI tools, so teams can securely analyze shared data with minimal setup. 
    
3. Extending Beyond ADB
   This isn’t just about Autonomous Database. Oracle’s Fusion Data Intelligence runs on ADB. With Delta Sharing, users can share operational data directly from Oracle Fusion Data Intelligence without modifying code or duplicating data. For example, a healthcare provider can send patient data from Oracle Fusion Data Intelligence to Databricks for AI-driven clinical trials, or to Tableau for clinical dashboards—securely and in real time.

How Oracle ADB and Delta Sharing Work Together

Oracle ADB supports bi-directional Delta Sharing—meaning it can act as both a data provider and a data recipient. ADB customers can share data with Databricks users or any platform that supports the Delta Sharing open protocol, and they can access data from these systems, all without duplicating or moving data manually.

Here are the 4 data sharing scenarios that this integration supports:

Let’s see how the data sharing works from ADB to Databricks.

Sharing from Oracle ADB to Databricks

ADB users can share operational or transactional datasets with Databricks to enable advanced analytics and ML workflows. This is done securely via Oracle’s Delta Sharing server, which allows access to datasets without physically moving the data.

1. Create a Share - An ADB administrator defines the dataset and creates a data share.
2. Grant Access - The recipient is sent a secure activation email with a downloadable credentials file (JSON).
3. Request Access - The recipient uses the credentials to authenticate and request the data.
4. Securely Access Data - The Delta Sharing server validates the request and returns a pre-authenticated request (PAR)—a short-lived, secure URL pointing to Parquet files in object storage.
5. Designed for Performance - Because data is read directly from object storage, the database server is not burdened with additional performance requirements.

This approach keeps things fast, secure, and scalable.

Here’s the demo that shows how to create a share in Oracle and consume that share in Databricks

For the rest of the scenarios, refer to Oracle documentation here.

KPMG: Delta Sharing eliminates silos and accelerates financial reconciliation

KPMG helps clients unify financial master data and transaction records across Oracle ADB and Databricks using Delta Sharing, eliminating redundant data movement and legacy integration patterns. A large national retailer is working with KPMG to modernize its financial reconciliation processes. Historically, the retailer’s financial master data and transactions were locked in a data warehouse, while reconciliation and reporting relied on disparate BI tools and custom integrations, leading to delays and data inconsistencies.

With this new integration, curated financial datasets and transaction views are securely exposed from Oracle ADB directly to BI tools and purchase-shipping reconciliation platforms using Delta Sharing. Its native connectors to Power BI, Tableau, and other analytics platforms enable financial analysts to access validated, real-time data for reconciliations—without the need for data extraction or replication. For IT, this approach simplifies architecture by removing legacy connection patterns, reducing maintenance overhead, and ensuring a single source of truth for financial reporting and analysis.

"Oracle and Databricks are important alliance partners in our partner ecosystem," notes Michael Juarez, Manager of Advisory Enterprise Analytics at KPMG. "Delta Sharing eliminates silos between databases, enabling quick implementation, native monitoring and traceability, and seamless integration with third-party BI tools."

The Road Ahead: Expanding the Partnership

The partnership between Oracle and Databricks continues to evolve with several exciting developments on the horizon:

• Change Data Feed Support - Oracle ADB will soon support Change Data Feed, allowing customers to share only the changed data. This will reduce unnecessary data movement and cut processing costs.
• Iceberg Table Support – Coming soon, customers will be able to use Apache Iceberg™ tables for sharing data. Iceberg is a modern table format that supports time travel, schema evolution, and high-performance querying, making it ideal for sharing large, dynamic datasets.
• Improved Authentication – Oracle ADB users can now authenticate into Databricks using their own identity provider (IdP) credentials, rather than relying on provider-issued OAuth tokens. This simplifies access management and strengthens security posture across both platforms.

Ready to experience the power of Delta Sharing between Oracle Autonomous Database and Databricks? Watch this video to learn how Oracle adopted Delta Sharing. Check out Oracle Live Labs, where you’ll find step-by-step guidance on setting up Delta Sharing between the platforms.

Don't miss the Data + AI Summit in San Francisco (Moscone Center, June 9–12)! Register and see the product and engineering team discuss the latest innovations in sessions like “What's New with Data Sharing and Collaboration with Live Demos” and “Delta Sharing in Action: Architecture and Best Practices”