Feature Highlights
AI Vector Search
Oracle AI Vector Search is designed for Artificial Intelligence (AI) workloads and allows you to query data based on semantics, rather than keywords.
JSON Relational Duality
Data can be transparently accessed and updated as either JSON documents or relational tables.
Developers benefit from the strengths of both, which are simpler and more powerful than Object Relational Mapping (ORM).
Operational Property Graphs in SQL
Developers can now build real-time graph analysis applications against operational data directly in the Oracle AI Database, utilizing its industry leading security, high availability and performance capabilities.
See Support for the ISO/IEC SQL Property Graph Queries (SQL/PGQ) Standard.
Microservice Support
Alongside Oracle's already comprehensive support for microservices, new functionality makes it simpler to implement cross-service transactions.
See Microservices.
Lock-Free Reservations
Lock-free column value reservations allow applications to reserve part of a value in a column without locking the row; for example, reserve part of a bank account balance or reserve an item in inventory without locking out all other operations on the bank account or item.
Kafka APIs for TxEventQ
Kafka applications can now run directly against the Oracle AI Database with minimal code changes, leveraging high performance Transaction Event Queues (TxEventQ).
JavaScript Stored Procedures
Developers can now create stored procedures using JavaScript in the database. This functionality also allows developers to leverage the huge number of JavaScript libraries.
See JavaScript.
Priority Transactions
Low priority transactions that block high priority transactions can be automatically aborted. This feature reduces the administrative burden on the DBA while maintaining high transaction throughput.
Data Use Case Domains
Data use case domains allow developers to declare the intended usage of data (columns) in a centralized and light-weight manner. For example, you can declare a column to hold an email, URL, password, currency, and so on. Applications can use data use case domains to automatically generate code or verify values.
Many Data Type and SQL Enhancements
The following features are among the many data type and SQL enhancements:
Improved Machine Learning Algorithms
New improvements to Oracle In-Database Machine Learning algorithms make it simpler to categorize text and data while offering better performance and flexibility.
Sharding Enhancements
New functionality makes it simpler to create and manage shard replicas. New sharding models also improve the distribution of data for shard keys with few unique values.
See Using Raft Replication in Oracle Globally Distributed Database.
Schema Privileges to Simplify Access Control
System privileges can now be granted at the schema level. This feature simplifies the privilege management process and as a result, makes it easy to secure databases.
Developer Role for Application Developers
A new role allows administrators to quickly assign developers only the privileges they need to design, build, and deploy applications for the Oracle AI Database.
See Use of the DB_DEVELOPER_ROLE Role for Application Developers.
Oracle SQL Firewall Included in Oracle AI Database
Included in Oracle AI Database, SQL Firewall provides real-time protection against common database attacks by monitoring and blocking unauthorized SQL and SQL injection attacks, no matter the SQL execution path.
Azure AD OAuth2 Integration
New functionality enables single sign-on to Oracle AI Database service instances or on-premises Oracle AI Databases from Microsoft Azure Cloud.
See JDBC Support for OAuth 2.0 Including OCI IAM and Azure AD.