The new Oracle AI Database 26ai marks a major milestone. With more than 300 new features, it focuses on artificial intelligence, developer productivity, security, and performance.
One of the highlights is AI Vector Search, which lets you leverage advanced AI models to analyze and compare text, images, or audio with impressive speed. Oracle also uses AI internally to optimize core database functions, such as estimating execution times and resource costs, making the engine even smarter and more efficient.
Developers now have greater flexibility: they can work with JSON or relational data or both together while the RAFT protocol makes it easier to distribute databases for performance or compliance requirements.
SQL and PL/SQL gain new data types and language enhancements, while SQL Firewall adds an extra layer of protection by monitoring and controlling which queries are executed. For DBAs, many management tasks have been simplified—space management, administration, and performance tuning with added boosts from features like True Cache, ensuring that queries run much faster than before.
Feature Highlights of 26ai
Here are some of the most important new capabilities:
- AI Vector Search: semantic search across text, images, and audio—not just keywords.
- JSON and relational data together: work seamlessly with JSON documents or relational tables, without the complexity of ORM.
- Real-time graph analytics in SQL: build graph-based applications directly on operational data with Oracle’s security and performance.
- Microservices and distributed transactions: easier implementation of cross-service transactions.
- Lock-free reservations: reserve part of a column value (like a fraction of a bank balance or an inventory item) without locking the entire row.
- Kafka and TxEventQ: run Kafka applications directly against Oracle Database with minimal changes.
- JavaScript stored procedures: write and run procedures in JavaScript directly inside the database, with access to its vast ecosystem of libraries.
- Priority transactions: critical transactions go first, while low-priority blockers can be automatically aborted.
- Data use case domains: declare how data should be used (email, URL, password, currency, etc.) to improve validation and code generation.
- SQL and data type enhancements: BOOLEAN support, Unicode 15.0, SELECT without FROM, simpler GROUP BY, and more.
- Machine learning improvements: faster, more flexible algorithms for classifying text and data.
- Simplified sharding: easier replica creation and new distribution models with RAFT.
- Simplified privileges: schema-level privileges and a new DB_DEVELOPER_ROLE for developers.
- SQL Firewall: built-in real-time protection against SQL injection and unauthorized queries.
Some of these features first appeared in earlier releases (19c, 21c, 23ai), but it’s with 26ai that they have truly matured. Now stable, enhanced, and fully integrated, they’re ready for production at scale.
This article highlights the key features of Oracle Database 26ai, but if you’d like to explore the full list in detail, you can check out the official documentation here:
Oracle Database 26ai – New Features
Poster un Commentaire