intro
Let’s explore what Oracle 23ai brings to the table and why it is one of the most exciting Oracle Database releases yet.
Oracle 19c still holds a large share of the market, but there’s no doubt Oracle is strongly encouraging migrations to Oracle 23ai. The reasons are a long support horizon (with Premier Support through 2031) and powerful new features for supporting AI applications.
In this blog post, you will learn what Oracle Database 23ai is, when it was released, and the main features and improvements it brings. By the end, you will have a solid understanding of whether upgrading to the latest version of Oracle makes sense for you.
Let’s dive in!
What Is Oracle Database 23ai?
Oracle Database 23ai, formerly known as Database 23c, is Oracle’s latest long-term support release (as of this writing). It has a Premier Support end date of December 31, 2031, while the Extended Support end date is still TBD.
Compared to previous Oracle database solutions, this is the first release that heavily focuses on features for AI. That is also why Oracle decided to rename it from Database 23c to the more technically catchy Database 23ai.
Oracle 23ai introduces a lot of new features, tweaks, and enhancements that aim to simplify data management and support AI-driven applications.
Oracle 23ai New Features
As explained in the announcement blog post, Oracle Database 23ai introduces over 300 new features. Listing them all here would be too lengthy and impractical.
Below, we will focus only on its most notable features.
Note: For a detailed breakdown, refer to the official Oracle 23ai New Features Guide documentation.
AI Vector Search
This is the flagship innovation introduced in Oracle Database 23ai. AI Vector Search allows the database to store and query vector representations of data, which is essential for capturing the semantic meaning of text, images, and other unstructured content. Learn more in our guide on vector databases.
This is made possible by the new VECTOR data type, along with dedicated indexes and SQL operators. Together, these enable fast similarity searches directly within the Oracle database.
Additionally, those new capabilities support RAG (Retrieval Augmented Generation) workflows, allowing LLMs to answer questions using your private business data without the need for external pipelines.
JSON Relational Duality Views
JSON Relational Duality Views bridges the gap between relational and document models. You can treat the same data either as JSON documents via document APIs or as relational tables using standard SQL.
This dual approach means you no longer have to pick between flexibility and performance. Instead, you can get both. In detail, you get the benefits of both the SQL and NoSQL worlds.
This feature makes it much easier to build applications that combine structured and semi-structured data, a common requirement in modern architectures.
Operational Property Graphs
Oracle 23ai comes with improved native support for property graph data models, allowing you to build and query graphs alongside transactional, JSON, and spatial data.
More specifically, it is the first commercial database to implement the new ANSI standard SQL/PGQ (Property Graph Queries). This language gives you the ability to express graph traversals and patterns using familiar SQL syntax.
The result is a simpler way to develop graph-based applications, such as fraud detection or network analysis, all within a single database.
SQL Firewall
Oracle Database 23ai includes a built-in SQL Firewall embedded directly into the database kernel. That means it can inspect all incoming connections and SQL statements.
SQL Firewall enforces user-specific policies, blocking unauthorized or suspicious SQL activity, including SQL injection attacks. You can configure it to allow, log, or block actions based on precise rules, adding a strong layer of protection to your data.
True Cache
True Cache makes caching in Oracle Database much easier. Instead of managing cache consistency yourself, you can deploy a diskless True Cache instance at the midtier. Then, Oracle’s Active Data Guard technology will guarantee that the cache always stays in sync with the primary database. This lowers the load on your database while achieving faster access to data.
RAC Scalability and Availability Improvements
Oracle is primarily an enterprise database. So, it is no surprise that high availability and scalability remain top priorities in Database 23ai. More in detail, this version introduces several important Oracle RAC (Real Application Clusters) enhancements:
Other Features and Improvements
Other oracle 23ai new features and enhancements worth mentioning are:
Further Reading
Expand your knowledge of Oracle 23ai with these selected resources:
Conclusion
In this blog post, you explored the world of Oracle 23ai, with a special focus on its new AI capabilities and other major improvements. Regardless of which Oracle version you plan to use, you will need a reliable database client, just like DbVisualizer.
DbVisualizer is a top-rated Oracle database client that also supports over 50 other databases. Its premium features include query optimization, drag-and-drop query building, ERD-style schema generation, streamlined import/export, and more. All that within an intuitive UI.
DbVisualizer is a robust, Java-based, enterprise-ready application trusted by leading companies worldwide. See how this tool can enhance your Oracle database experience by trying DbVisualizer Pro for 21 days!
FAQ
What is the difference between Oracle Database 23c and Oracle Database 23ai?
There is no actual difference between the two. Given the significance of the AI innovations in this release, Oracle decided to rename Database 23c to Database 23ai. Specifically, Oracle Database 23ai corresponds to version 23.4.0.0.0 and will be known by this new name going forward.
Where is it possible to download Oracle Database 23ai?
Oracle Database 23ai Free can be downloaded from the official site as a container image, available for Linux, macOS, and Windows.
When was Oracle 23ai release date?
Oracle Database 23ai was released on May 2, 2024. It initially became available on OCI through services like Exadata Database Service, Exadata Database Cloud@Customer, and Base Database Service.
When was Oracle 23ai on prem release date?
The on-premises version was also released on May 2, 2024, alongside the cloud version.
Why should you upgrade from Oracle Database 19c to 23ai?
Oracle Database 23ai is the next long-term support release after 19c, offering five years of Premier Support plus three years of extended support. While 19c’s Premier Support ends in 2029 (and Extended Support in 2032), upgrading now guarantees longer coverage, access to new AI features, and continued security updates.
Is there a free trial available for Oracle Database 23ai?
Yes, Oracle offers a free edition called Database 23ai Free. This version is available for anyone to download and use, and is ideal for developers, data scientists, or anyone who wants to explore the latest features of Oracle Database for AI. Alternatively, you can try the premium version of Oracle Database 23ai through the OCI (Oracle Cloud Infrastructure) by using the Autonomous Database Free Tier.
Which LLMs and embedding models does AI Vector Search work with?
AI Vector Search has been officially tested with GPT-3.5, GPT-4, Cohere, Llama2, Mistral, Gemini, Palm2, and Vertex AI. Note that many of these LLMs are available on Oracle Cloud. It also supports over 90 embedding models and integrates with popular frameworks like LangChain and LlamaIndex.
