Announcing Expanded Integration Between Oracle Database and the Snowflake AI Data Cloud

What if the very events that run your business — the transactions clearing in retail banking, the swipes at every point of sale, the movements across global supply chains, the records changing inside Oracle — could immediately fuel intelligent action? With the Snowflake Openflow Connector for Oracle Database, those events would no longer remain locked inside operational systems. They could flow securely into Snowflake, where Snowflake Intelligence, powered by Snowflake Cortex Agents (both currently in public preview), would analyze them alongside every other enterprise signal to reason, decide and act in real time. For industries that rely on Oracle, such as banking, retail, life sciences and healthcare, this would transform static operations into live intelligence.

At Oracle AI World, we announced expanded integration options for joint customers and the private preview release of the Snowflake Openflow Connector for Oracle, a fully built-in connector jointly engineered for simplicity, reliability and near real-time data movement. This collaboration brings together Oracle’s established database technology and Snowflake’s scalable AI Data Cloud, allowing organizations to stream operational data into Snowflake directly as new transactions occur. Powered by Oracle’s log-based change data capture (CDC) and Snowflake’s native streaming capabilities, the solution removes pipeline complexity, enables low-latency insights and supports a range of deployment environments.

Whether supporting AI, analytics or regulatory reporting, the Oracle and Snowflake integration powers enterprises to unlock agility and insights at scale, backed by the deep expertise and unified support from both teams.

Two powerful choices for a proven architecture

We recognize that different customers have different needs, so we are providing two methods for near real-time data replication. By providing choice, we aim to enable every organization to implement the optimal data flow for its specific requirements.

1. Snowflake Openflow Connector for Oracle for a streamlined, native experience

To provide an efficient, fully managed experience within Snowflake, we are excited to introduce the Snowflake Connector for Oracle. Built on the Snowflake Openflow framework, the connector is designed for simplicity, performance and operational efficiency.

• Streamlined and native experience: Built for ease of use, the connector offers a zero-installation experience directly within the Snowflake platform, reducing operational overhead for data teams.

• Fully managed and scalable capabilities: As a built-in Snowflake component, the connector is managed by Snowflake. It automatically scales to handle data volumes, freeing engineers from managing complex pipeline infrastructure.

• Flexible licensing: Customers can choose between an embedded, Snowflake-managed license or a bring-your-own-license (BYOL) option for those who have an existing Oracle GoldenGate license.

2. GoldenGate for control, streaming processing and cloud innovation

Alternatively, Oracle offers Oracle Cloud Infrastructure (OCI) GoldenGate, which sets the standard for enterprise replication and data integration.

• Enterprise-grade control: Delivers fine-grained controls, high performance integration with Oracle SGA (System Global Area) transactions and broad heterogeneous replication with hundreds of third-party data stores.

• Innovation for Snowflake: Oracle supports streaming replication for Oracle Database via Snowpipe Streaming API for low-latency loading. GoldenGate also provides a Stage-Merge (micro-batch) solution that handles merge operations from complex DML/DDL in a flexible and low-cost manner.

• Stream processing and analytics: Is available in the GoldenGate platform to handle streaming extract, transform, load (ETL), data cleansing and time-series analytics on data events that are in-flight on their way into the Snowflake data warehouse and lakehouse systems.

A deep dive into the Openflow Connector's architecture

Our design philosophy for the Openflow Connector for Oracle was centered on three pillars: enterprise-grade performance, operational simplicity and native Snowflake integration.

At its core, the connector is built on Oracle’s native XStream API. This is a deliberate and critical design choice. The XStream API shares some of the same powerful technology that runs inside of Oracle GoldenGate, allowing Snowflake to tap directly into the database's redo logs. This log-based CDC captures committed inserts, updates and deletes with minimal impact on online transaction processing (OLTP) systems, a clear advantage over more trigger-based, query-based and logminer-based CDC tools.

We prioritized a zero-footprint, agentless architecture. There are no agents to install, manage or patch on Oracle servers. The setup on the Oracle side is accomplished by a database administrator running a standard set of SQL commands to configure and enable the XStream outbound server. From there, all integrations are controlled in Snowflake. Using the Openflow UI, users can configure replication scope (for example, schemas, tables and columns); monitor pipeline health with telemetry and alerts; and track ingestion in near real time.

As shown in Figure 1, the Openflow Connector reads logical change records (LCRs) directly from the XStream outbound server queue in near real time. These change events are then streamed immediately into Snowflake target tables using Snowpipe Streaming. This direct, memory-to-memory pipeline avoids landing data in intermediate files, resulting in fewer hops and points of failure, and delivers end-to-end latency in seconds.

For unique requirements, the replication flow is built on the flexible Openflow framework. Teams can open the Apache NiFi canvas to expand pipelines. For example, teams can enrich data with lookups or mask sensitive PII data before it reaches Snowflake.

This flexible architecture supports Oracle Database versions 12c R2 (12.2), 18c, 19c, 21c, 23ai and 26ai running on-premises, on Oracle Exadata, in OCI (VM/Bare Metal) and on AWS RDS Custom for Oracle.

Trusted partnership, simplified support

Connecting mission-critical systems requires deep trust and reliability. The Snowflake Openflow Connector for Oracle is a product of close engineering collaboration between Snowflake and Oracle, with a streamlined support model. Customers engage Snowflake as the single support point, while an Oracle partner helps ensure direct support for Oracle components when needed. This removes the need for superfluous handoffs and guarantees coordinated resolution.

For Snowflake, this unlocks a critical enterprise data source for the AI Data Cloud. For Oracle, it extends the reach of its databases into modern cloud and AI ecosystems. For customers, it provides a best-of-breed architecture backed by two industry leaders committed to success.

From transactional data to AI-powered insights

The most valuable AI applications are built on trusted enterprise data. For decades, Oracle databases have powered mission-critical transactional and operational workloads with unmatched performance and integrity.

The challenge has been to make this high-fidelity data available in real time to drive modern analytics and intelligent cloud applications. The Snowflake Openflow Connector for Oracle creates a seamless, near real-time bridge between Oracle systems and the Snowflake AI Data Cloud. This lets you extend the value of your Oracle data beyond operations to fuel predictive insights and next-generation applications, all without impacting source system performance. Let’s take a closer look at some scenarios for each industry: 

1. Banking, financial services and insurance for fraud detection

Consider a retail bank processing millions of credit card transactions hourly, all recorded in a high-throughput Oracle OLTP system. Their existing fraud detection models run in batches, catching threats too late. By streaming every transaction into Snowflake within seconds via the Openflow Connector or GoldenGate, the bank feeds live data into Snowpark ML models for real-time fraud detection. Suspicious transactions are flagged and blocked before they are completed. Additionally, Snowflake Cortex AI's anomaly detection can add agility to identify fraud patterns faster than legacy batch systems.

2. Retail for dynamic inventory and personalization

A global retailer relies on a mission-critical Oracle database that powers its core inventory and order management systems. But this critical data is only loaded into its analytics platform overnight. With Openflow Connector for Oracle or GoldenGate streaming replication, the company gets one consistent up-to-date view of the entire business operation. This unlocks AI use cases. Forecasting through Cortex AI can predict item demand by region to trigger restocking, while Snowpark models update customer profiles and push personalized offers to boost revenue as purchases happen.

3. Manufacturing for predictive maintenance and supply chain optimization

A global automotive manufacturer runs its production line and supply chain on Oracle databases, but latency in getting that data to analysts means a critical equipment failure can halt production for hours. By streaming sensor and logistics data into Snowflake using the Openflow Connector or Oracle, they create a live view of their entire operation. This allows a Snowpark ML model to predict equipment failures before they happen, avoiding downtime. Simultaneously, a Cortex AI model monitors the live supply chain data, automatically rerouting materials to mitigate disruptions and keep the assembly line running.

Get started today

The Openflow Connector for Oracle is now available in private preview with public preview coming soon. To learn more and get started, we invite you to:

• Read the technical documentation and Setup Guide.

• Contact your Snowflake account team to get access to the connector. 

The launch of Snowflake Openflow Connector for Oracle is designed to reduce complexity and increase the flexibility of your enterprise data integration workflows. We are excited to see the incredible applications you will build on this foundation of trusted enterprise data.

Forward-looking statements

This article contains forward-looking statements, including about our future product offerings, and are not commitments to deliver any product offerings. Actual results and offerings may differ and are subject to known and unknown risk and uncertainties. See our latest 10-Q for more information.