Announcing GoldenGate 23ai General Availability

April 19, 2024 | 8 minute read
Peter Inzana
Product Manager, GoldenGate
Text Size 100%:

Oracle GoldenGate announces its latest long-term support release, GoldenGate 23ai, and makes it available on premises and within OCI GoldenGate. Oracle GoldenGate is the market leading technology for real-time heterogeneous data integration and building global transactional data fabrics.  Among its many new features, it now supports real-time replication for enterprise-class vector embeddings to improve Generative AI and large language model accuracy. 

GoldenGate 23ai supports new Oracle Database 23ai features in addition to over a hundred updates to the GoldenGate portfolio that improve usability, performance, security, governance, and developer APIs. GoldenGate 23ai features new capabilities that allow real-time replication of Oracle vector data types and across non-Oracle vector databases. This is especially important for distributed applications and data where unifying vector embeddings is important to optimize LLM accuracy.  It also allows enterprises to quickly bring AI to all their data with minimal risk by replicating data from their existing databases to Oracle Database 23ai where it can be vectorized and indexed for fast AI search.

Watch a video overview of the new features. 

Oracle GoldenGate 23ai new feature summary

Organizations are struggling with maintaining consistency, resiliency, and security across modern data environments which are characterized as being highly distributed, diverse, and dynamic, " said Stewart Bond, Research VP, Data Intelligence and Integration Software at IDC. "Oracle GoldenGate 23ai is Oracle’s next major release that adds 100+ GoldenGate enhancements that improve usability, performance, resiliency, and security. More specifically, GoldenGate 23ai's new user experiences, support for vector databases, and Data Streams implemented through the AsyncAPI directly addresses modern data environment demands for the modern data worker.

GoldenGate 23ai delivers new, advanced capabilities across the GoldenGate platform that enable heterogeneous, hybrid, and multicloud data synchronization for operational, analytical, and distributed GenAI applications.   

This blog highlights the following major new features:

  • Data replication for AI vectors and vector databases
  • Data Streams for Pub/Sub access to database events via AsyncAPI
  • Updated graphical user experience
  • Improved observability and diagnostics
  • Oracle Database 23ai new feature support
  • Enhanced security management and user authorization
  • Extended heterogeneity

Data replication for AI vectors and vector databases

Transforming unstructured data into vectors is fast becoming the mission for business everywhere. GoldenGate 23ai is the means to bring your real-time unstructured data and vectors to your AI applications where you can deliver next generation user experiences.

Chapter 1

Oracle GoldenGate 23ai adds replication support for the Oracle Database 23ai vector data type which can power enterprise RAG (Retrieval Augmented Generation) and consolidate distributed vector and data into a vector hub for real-time GenAI applications experiences. Ensuring replication of trusted enterprise transactional data into LLMs (Large Language Models) in real-time is critical for improving GenAI usefulness (a combination of effectiveness and timeliness). GoldenGate 23ai can stream “AI ready” data into GPTs (Generative Pre-trained Transformers) for fewer hallucinations and more accurate RAG (Retrieval-Augmented Generation) results.

GoldenGate 23ai adds new vector databases to its connectivity portfolio including Oracle Database 23ai, MySQL Heatwave vectors, PostgreSQL (pgvector), GCP AlloyDB AI, AWS Aurora (pgvector), AWS RDS PostgreSQL (pgvector), Azure PostgreSQL (pgvector), ElasticSearch vectors, OpenSearch vectors, AWS OpenSearch vectors and OCI OpenSearch Neural.Vector Databases

In this blog, you can learn more about how GoldenGate 23ai can migrate vectors into Oracle Database 23ai.  Here are four example patterns using GoldenGate 23ai for distributed vector processing:

Figure 1:  Four use cases for managing real time distributed vectors

Figure 1:  Four use cases for managing real-time distributed vectors

GoldenGate 23ai enables real-time unstructured data and derived vectors to move across your distributed data fabric.  A popular use case is to build a vector hub where information is consolidated and supports fine tuning vector searches with private, secure, and trusted enterprise information. 

Figure 2:  Popular use case for uses multi-data source vector replication

Figure 2:  Popular use case for uses multi-data source vector replication

Data Streams for Pub/Sub access to database events via AsyncAPI

Event-based architectures enable for more dynamic and responsive applications.  Data Streams is a new capability that allows API access to all GoldenGate fault tolerant replication, including database inserts, updates, and deletes as well as database schema changes such as field attributes. 

Chapter 2

With direct API access, data fabrics can be even more responsive.  Applications can publish fault-tolerant messages rather than just Kafka topics, and data subscribers can reduce latency even further in publishing messages directly between microservices.  Learn more about the AsyncAPI standard used by Oracle GoldenGate.  Listen to an architectural overview of Data Streams and the AsyncAPI implementation.

With a pub/sub approach, this new capability exposes database (DDL and DML) events, from any of the vast array of databases supported by GoldenGate’s microservices-based transaction log capture methods, directly via the AsyncAPI channels, as JSON formatted CloudEvents.

Figure 3:  Data Streams outputs all DDL/DMLS database events for pub/sub usage

Figure 3:  Data Streams outputs all DDL/DMLS database events for pub/sub usage 

New graphical user experience

To enable a more productive user experience, GoldenGate 23ai simplifies setup and administration with new user experiences for workflows, configuration, management, and troubleshooting. 

Chapter 3

GoldenGate 23ai updates graphical user experiences for building replication and continuous transform and load (CTL) pipelines, enabling more automated workflows and visibility to operational data.

Chapter 4

Figure 4:  New GoldenGate 23ai user interfaces for Administration, Distribution/Receiver and Performance Metrics services.

All replication workflows have been streamlined for more efficient configuration and tuning. GoldenGate process information retrieved from the Oracle UTL_RAPDV and AWR reports is now visible in the GUIs, allowing for more advanced troubleshooting from a single location.  Visibility into large numbers of concurrent Extract and Replicat processes has also been improved.

Improved observability and diagnostics


GoldenGate 23ai also features significant enhancements for observability and diagnostics, providing a more comprehensive and intuitive approach for monitoring all processes, status and performance of replication flows across multicloud topologies. 

GoldenGate optionally pushes telemetry in StatsD format for consumption by customers or telegraf.  All metrics collected can be streamed for real-time alerts, using GoldenGate Stream Analytics or Kafka, and all are stored in an Oracle Database for historical analysis.

Learn more in this blog.

Oracle Database 23ai new feature support

Oracle Database 23ai new feature support GoldenGate 23ai is fully integrated and optimized for usage with the Oracle 23ai Database.

Chapter 4

Database 23ai adds database features that require replication support as well as improved observability and diagnostics during integrated operations.

Database 23ai supported features

As always, GoldenGate 23ai supports all Oracle Databases, including Autonomous Database, Exadata Cloud Service, Exadata Cloud at Customer, and Base Database service, as well as on premises software versions. And together, the latest version, 23ai, of both GoldenGate, 23ai the Oracle database, provide combined power and unique differentiation no other vendor can match for enabling global, distributed, transactional and analytical data fabrics.

OCI GoldenGate is the best, low-impact technology to capture changed data from Autonomous Database for replication to any target or consuming application.


Enhanced Security management and user authorization

Chapter 5

GoldenGate 23ai adds several, enterprise-strength, multi-form advancements in security. These new features include encryption of data at rest and in-flight, role-based authentication and privileges, and integration with multiple identity providers.    


Extended Heterogeneity

GoldenGate enables data-driven innovation regardless of whether it’s deployed on-premises, in a 3rd party Cloud and/or in Oracle Cloud Infrastructure (OCI). However, leveraging OCI GoldenGate 23ai provides the best price-performance, across a multitude of deployment options, spanning real-time heterogeneous data replication, stream analytics, and now even batch/bulk “Data Transforms” that extend OCI GoldenGate’s abilities for complex transformations, its reach to more enterprise SaaS applications, as well as other sources and targets.  

Figure 8

In Summary

GoldenGate 23ai supports new Oracle Database 23ai features and move than one hundred additional enhancements to the GoldenGate portfolio that improve usability, performance, security, governance, and developer APIs.

GoldenGate 23ai is now supported in OCI GoldenGate and is available in 45+ world-wide data centers as fully-managed OCI cloud service with elastic auto-scaling for optimized performance with FinOps control.  OCI GoldenGate is used by thousands of customers for these use cases:  cross-region/cross-cloud high availability, high-speed ingest to any cloud data warehouse/lakehouse, ground-to-cloud data migrations, and analyzing and enriching data in-motion.

Get started today!

Try GoldenGate 23ai on premises, in the cloud, and in OCI GoldenGate in your existing tenancy or open an OCI Free Tier account.

Peter Inzana

Product Manager, GoldenGate

Peter Inzana is a Director in Oracle GoldenGate Product Management and is a key spokesperson.  He eagerly and expertly promotes the solutions with customers, analysts, and partners.  Peter has more than 20 years of experience in data integration, replication, and governance.

Previous Post

OCI GoldenGate User Assistance: Find your content faster with filters!

Katherine Wardhana | 3 min read

Next Post

OCI GoldenGate: Capture a topic data from Confluent Cloud to Autonomous Database

Madhu Kumar S | 12 min read