Changes in This Release for Oracle Deep Data Security

Learn about the new features available for Oracle Deep Data Security in Oracle AI Database 23.26.4.

Feature	Description
Data grant filters	You can create data grant filters that further limit the rows and column values available through data grants. See Create Data Grants.
SQL macros in data grants	You can use SQL macros to reuse predicate logic across multiple data grants. See Use SQL Macros in Data Grant Predicates.
Configurable OAuth 2.0 provider	You can integrate OAuth 2.0 providers using the new OAUTH2 identity provider type and configurable token validation settings. Keycloak is the initial supported provider. See Configure Keycloak for Application-Mediated Access and Set Up IAM Integration for Application-Mediated Connections.
Optional end-user token and application identity–based connection pools	You can omit the separate end-user token from an end-user security context payload and use the database-access token for authentication and authorization. With an application identity–based connection pool, the database-access token in the payload can identify a different application from the one that established the connection. See About the End-User Security Context.
Audit enhancements	Standard database auditing now captures all end-user security context details. You can view these details through the UNIFIED_AUDIT_TRAIL view. See UNIFIED_AUDIT_TRAIL in Oracle AI Database Reference.
Data authorization for analytics	A new chapter explains how to enforce data authorization using Oracle Deep Data Security for analytical SQL. See Enforce Data Authorization for Analytics.