New Features in 23.5 Release Update

This section describes the new features in the 23.5 release update.

Archiving and Unarchiving of Gold Images

You can archive and unarchive Oracle FPP gold images that are not currently used but cannot be deleted for future access. Archiving and unarchiving gold images allow you to store those images on external storage devices of your choice in a space-efficient, compressed fashion, thereby freeing up storage space on high-end storage devices hosting the central FPP gold image repository.

Archiving gold images that are not currently used, but need to be retained, saves space and costs by allowing you to flexibly and efficiently store them on external storage devices of your choice.

View Documentation

BINARY Vector Dimension Format

BINARY is a new dimension format that can be used with the VECTOR data type. Each dimension of a BINARY vector can be represented with a single bit (0 or 1). A BINARY vector itself is represented as a packed UINT8 array, for example, a single UINT8 value represents 8 dimensions of the BINARY vector. BINARY vectors can be generated using embedding models provided by Cohere (for example, embed v3), Hugging Face Sentence Transformers, and so on.

BINARY vectors offer two key benefits compared to FLOAT32 vectors:

  1. The storage footprint of BINARY vectors is 32X lesser, and
  2. Distance computations on BINARY vectors can be up to 40X faster, which accelerates Vector Search

BINARY vectors can provided reduced accuracy compared to FLOAT32 vectors. But, evaluations on various datasets have shown that they can still achieve 90% or higher accuracy of FLOAT32 vectors.

View Documentation

Backup, Restore, and Relocation for FPP Server

You can create a backup of the Oracle Fleet Patching and Provisioning (Oracle FPP) server, restore data from the backup, and relocate the server from backup to new hardware. Relocate the Oracle FPP Server to new hardware and re-point Oracle FPP targets whenever needed.

Ensure data safety by backing up Oracle FPP Server and restoring data in case of failure.

View Documentation

Custom Certificates for Oracle FPP Server Authentication

Starting with Oracle Grid Infrastructure 23ai, you can specify custom security certificates for authentication between Oracle Fleet Patching and Provisioning (Oracle FPP) server and clients. By default, Oracle FPP uses SSL/TLS certificates. You can use any security certificate assigned by a Certificate Authority of your choice.

Choosing custom security certificates enables you to meet stringent security policies and regulations set by your organization.

View Documentation

DB Operator RAC Controller on Kubernetes

Integrate and operate containerized Oracle Real Application Clusters (Oracle RAC) databases on Kubernetes. Together with Helm Charts (a package manager for template YAML files and other capabilities), this feature provides a simple, stateless, and integrated management framework to define, install, deploy, and delete Oracle Real Application Clusters deployments.

Deploy your Oracle Real Application Clusters databases conveniently in Kubernetes clusters.

View Documentation

Duplicated HNSW Vector Indexes on RAC

HNSW Vector Indexes are now supported on RAC environments through full duplication on all instances of the cluster that have  sufficient memory in the Vector Pool. On Autonomous Database Serverless deployments, the Vector Pool is autonomously managed.

All copies of the HNSW index across different RAC instances share the same ROWID-to-VID mapping table on disk. However, each instance builds its in-memory neighbor graph independently, and hence, its possible to get different results for approximate searches depending on which RAC instance is used to serve the query.

Enterprise customers often deploy Oracle Database in RAC environments. This feature enables creation of HNSW vector indexes for RAC through full duplication across all instances of the cluster. Queries directed at any instance of the RAC cluster can take advantage of HNSW vector index plans resulting in ultra-fast similarity searches.

View Documentation

General Improvements for Oracle FPP Job Scheduler

The Oracle Fleet Patching and Provisioning (Oracle FPP) job scheduler now supports pausing and resuming jobs, forcing jobs to start in a paused state, and pausing jobs between batches in a chain execution and allows to tag jobs for easy querying.

Additional job scheduler options provide you with more control over the patching execution. You can pause jobs to verify executed jobs and resolve dependencies before resuming jobs.

View Documentation

General Purpose Cluster Configuration

The General Purpose Cluster is a new cluster deployment that requires minimum network and storage configuration and supports applications that benefit from managed server restarts and failover capabilities. Oracle Grid Infrastructure installer provides a streamlined option to configure a general-purpose cluster in either interactive or silent mode. 

This feature increases productivity and reduces the required deployment time by simplifying deployment requirements. 

View Documentation

Gold Image Based Out of Place Patching

Oracle DBCA can apply Release Updates (RUs) out-of-place using Gold Images for both Oracle Grid Infrastructure homes and Oracle Database homes.

Oracle DBCA can apply the new Gold Image-based quarterly Release Update patches to Oracle Grid Infrastructure home and Oracle Database homes.

View Documentation

JSON Collections

JSON collections are a special table or view. Similar to Duality views it has just one column (called data) to hold a JSON document (as JSON type). A document is identified by an ID value. JSON collection tables|views are meant to simplify SQL access and be fully interoperable with form SQL. For example, it is possible to do a simple INSERT AS SELECT into a JSON collection table - if id values are not present in the JSON document they will be injected automatically. JSON collection tables and views are MongoDB compatible and also interoperable with JSON Duality view - in fact, a JSON Duality view is also a JSON collection view.

JSON Collection tables|views can be seen as a replacement for SODA collections which were not first class citizens in the database and therefore harder to use with SQL.

Native JSON collections simplify working with JSON data in Oracle Database. They make it easier to generate reports as JSON documents from SQL queries and expose JSON document sets to document-centric APIs such as Oracle Database API for MongoDB. 

View Documentation

JSON_ID SQL Operator

SQL operator JSON_ID generates a unique document-identifier value, for unique access to JSON documents in a collection. The argument to JSON_ID determines whether the value is a 12-byte OID or a 16-byte UUID. In Oracle JSON collections, JSON_ID is used to create (automatically or explicitly) the values for document-identifier field _id.

JSON_ID simplifies the generation of ID values to uniquely identify JSON document.

View Documentation

Optimized Oracle Native Access to NVMe Devices Over Fabric

Starting with Oracle Database 23ai, you can use TCP/IP network connections to access the remote NVMe storage devices using NVMe over Fabrics (NVMe-oF). The Oracle Grid Infrastructure server works as an initiator that connects to an NVMe-oF storage target created using Linux Kernel nvmet_tcp module,  providing optimized user mode access to remote NVMe devices.

Starting with Oracle Database 23ai, you can use TCP/IP network connections to access remote NVMe storage devices using NVMe over Fabrics (NVMe-oF). The Oracle Grid Infrastructure server works as an initiator that connects to an NVMe-oF storage target created using the Linux Kernel nvmet_tcp module, providing optimized user mode access to remote NVMe devices.

NVMe-oF provides a low-latency and secure way to access remote NVMe devices exported using NVMe Over Fabrics target. Oracle provides an optimized way to access these NVMe-oF devices directly from an Oracle process. This Oracle-native method of accessing NVMe-oF devices reduces latency while Oracle ASM makes storage manageability easier.

View Documentation

Oracle DBCA Support for PMEM Storage

Oracle Database Configuration Assistant (Oracle DBCA) enables you to select persistent memory database (PMEM) as your storage option when creating a single-instance database.

This feature automates the process of assigning a PMEM device for your database storage enabling you to place database files in a PMEM storage device.

View Documentation

Oracle DBCA Support for Standard Edition High Availability

Using the Oracle Database Configuration Assistant (Oracle DBCA) and facilitating Oracle's Automatic Storage Management or Oracle's Advanced Cluster File System, you can now quickly create a Standard Edition High Availability Oracle Database fully configured for automatic failover. 

Oracle Standard Edition High Availability Database can now be created very easily with more automation, eliminating manual steps and the associated complexity.

View Documentation

Oracle Database Installer Command-Line Support

Oracle Database Installer now supports specifying commands and input parameters for those commands using the command-line interface.

Easier and simpler Oracle Database deployments are supported using the command-line interface in addition to the graphical user interface.

View Documentation

Oracle FPP Lite Without Java Container

Oracle Fleet Patching and Provisioning (Oracle FPP) Lite formerly known as FPP local mode, does not require the Grid Infrastructure Management Repository (GIMR) and the Java container.

Oracle Grid Infrastructure and Oracle Database administrators can use Oracle FPP Lite without setting up any additional components on a cluster, making the patching process simpler and faster.

View Documentation

Oracle FPP Metadata on External Databases

Starting with Oracle Grid Infrastructure 23ai, Oracle Fleet Patching and Provisioning (FPP) stores metadata in a local Standard Edition High Availability database or an external Oracle metadata repository. New installations will store the metadata by default in a SEHA database.

Allowing to choose an external metadata repository database or a local Standard Edition High Availability (SEHA) database simplifies the deployment of the Fleet Patching and Provisioning server.

View Documentation

Oracle Grid Infrastructure Installer Command-Line Support

Oracle Grid Infrastructure Installer now supports specifying life-cycle management operations and input parameters for those operations on the command line.

Easier and simpler Oracle Grid Infrastructure deployments are supported using the command line, in addition to the graphical user interface.

View Documentation

Oracle Grid Infrastructure Installer Improvements

The Oracle Grid Infrastructure installer has been upgraded with options to create and manage gold images and perform out-of-place patching while reducing the inventory metadata to effectively manage installation and patching.

Out-of-place patching using the Oracle Grid Infrastructure Installer directly makes patching more manageable and reliable.

View Documentation

Single-Server Rolling Database Maintenance

Single-Server Rolling Database Maintenance creates a new local database home and starts a second instance of the same database from the new home on the same server, allowing you to perform rolling patching and maintenance operations on a single server hosting an Oracle RAC One Node or Real Application Clusters (Oracle RAC) database.

Single-Server Rolling Database Maintenance provides database availability during maintenance activities (such as patching) on a single server hosting an Oracle RAC or Oracle RAC One Node database. This feature significantly improves the availability of your single-node databases without expanding them to a multi-node cluster and adding support for shared storage.

View Documentation

Store Images and Transfer Working Copies as ZIP Files

Store Oracle Fleet Patching and Provisioning (Oracle FPP) gold images as ZIP files, create ZIP files from existing Oracle homes, and transfer these ZIP files.

Save significant storage, bandwidth, and transfer time by storing and transferring gold images as ZIP files. 

View Documentation

Vector Memory Pool Automatic Management

When using Autonomous Database services (ADB), the Vector Pool dynamically grows and shrinks when HNSW indexes are created or dropped respectively.

Because you cannot explicitly set any SGA-related memory parameters when using Autonomous Database services, this feature automatically maintains the necessary amount of Vector Pool memory needed for your HNSW indexes.

View Documentation

Verifying Digital Signature and Integrity of Installation Archive Files

Starting with Oracle Database 23ai, Oracle digitally signs the installation archive files with Oracle certificates.

Oracle digitally signs the installation archive files to allow customers to ensure the integrity of the packages before deploying them in their environments.

View Documentation