Oracle Database 23ai, from the version suffix "g", "c", to today's "ai", is enough to show Oracle's confidence and determination in the AI era. From the instant hit of ChatGPT to the stunning debut of Sora, generative AI has opened people's eyes to the infinite possibilities of the intelligent world, and this leapfrog innovative application has not only changed the way of interaction and creation of life and entertainment, but also allowed enterprises to see more opportunities to reshape business value, whether it is application development, configuration and deployment, system operation and maintenance, or business sustainability, all of which have been revitalized with the blessing of AI technology. Behind all this, data has always been the cornerstone of innovation. That's what Oracle does best.
"It's not so much about how many big language models you have, it's about how to use them. With a focus on AI for Data, Oracle Database has embedded generative AI into our enterprise applications, including Oracle Fusion Cloud Applications, NetSuite, and industry applications such as Cerner. "Many of our customers put their business-critical data on Oracle databases, and for us, the only criterion for using generative AI well is what it does for our customers." ”
Oracle invests an average of $6 billion annually in R&D, laying a solid foundation for AI-based technology innovation. The long-term support for Oracle Database 23ai includes Oracle AI Vector Search and more than 300 new key features to help simplify the use of AI in data, accelerate application development, and run mission-critical workloads. For example, customers can take advantage of the new AI Vector Search feature to securely search documents, images, and other unstructured data in combination with private business data without having to move or copy the data.
With the complexity of data types and data specifications, enterprises are faced with more and more diverse database choices, and customers often want to solve data problems in the database, rather than dealing with them at the application level. If the problems of the data layer are simplified and all the problems are piled up at the application layer to solve, then after the arrival of generative AI, the processing of the application layer will be more exhausted.
Because Oracle Database 23ai integrates AI Vector Search, you can integrate business data and vector data, directly expand vector columns, and search them through SQL, eliminating the need for data copying and greatly improving business efficiency. Customers can easily search for documents, images, and relational data based on conceptual content rather than specific text, pixels, or data values. AI Vector Search enables LLMs to query private business data using a natural language interface and helps LLMs deliver more accurate and relevant results. With AI Vector Search, developers can easily add semantic search capabilities to new and existing apps. By storing and processing business and vector data in the same high-performance database, customers can seamlessly integrate AI Vector Search into existing business applications to enable new AI use cases without compromising data security.
Oracle Database 23ai可以将AI算法引入到数据所在的位置，而不必将数据迁移到AI算法所在的位置。 因此，AI可以在Oracle数据库中实时运行，提高AI的有效性、效率和安全性。 展开来说，Oracle Database 23ai的AI特性体现在三个方面，一是面向数据的AI，包括Oracle AI Vector Search、Oracle Exadata System Software 24ai、OCI GoldenGate 23ai；二是加速应用开发，包括JSON Relational Unification、Graph Relational Unification、免费开发人员数据库；三是关键任务数据，包括Oracle Globally Distributed Database with RAFT、Oracle True Cache、In-Database SQL Firewall，可以在数据库融合AI能力，无需把业务数据调出交给第三方处理。
"Oracle Database 23ai focuses on AI, and pays more attention to how AI can be quickly integrated into enterprise applications and can be truly implemented. We have enabled our customers to add AI technology to their core business systems to unleash more capabilities, benefiting not only existing Oracle customers, but also many new customers, and our converged database has significant advantages over the combination and skill requirements of a single database for different scenarios. Li Jia, senior director of Oracle's China Technology Consulting Department, said.
Jia Li, Senior Director of Technical Consulting, Oracle, China
Through the concept of Data for Dev App, Oracle integrates relational data and object data, and can design any number of JSON views on the table of relational data, so that JSON developers no longer need to use SQL, and SQL developers can directly use JSON mode to access and use the entire database, and obtain SQL applications, which are suitable for a variety of knowledge bases. For some complex applications, Oracle's Lock-free can unlock data activity at the time of data activity and lock it when it is submitted, ensuring that long-running transaction performance is not affected, and at the same time, it will also sort transactions according to their priority.
In addition, Oracle has made mission-critical system architectures simpler and easier to scale, including support for RAFT in globally distributed databases, to meet various business requirements such as enterprises going global. As an in-memory, consistent, application-transparent, high-performance middle-tier cache, Oracle True Cache improves application response times and reduces the load on database servers. Unlike other middle-tier caches, True Cache data automatically maintains transactional consistency at every point in time. True Cache doesn't require developers to write code to populate and manage the data in the cache, making it easier to deploy between applications and databases. True Cache provides all Oracle SQL, JSON, and Graph query functions, as well as SQL firewall in databases, real-time SQL plan management, and other functions.
With the new SQL embedding function, customers can deploy models that meet the standards in Oracle's database, and with the help of the latter's CPU database engine for vectorization, users only need to master SQL to meet the requirements of AI development, realize SQL embedding, vectorization, search and display, and make AI technology more quickly implemented in enterprise applications. OCI GoldenGate 23ai supports cross-cloud data storage for heterogeneous data integration and high availability. GoldenGate 23ai provides a new feature that supports real-time replication of vectors between heterogeneous vector stores. GoldenGate23ai also enables enterprises to replicate data from existing databases to Oracle Database 23ai, quickly ingesting AI into all data in a low-risk manner, vectorizing and indexing it for fast AI searches.
据了解，Oracle Database 23ai会在基于Oracle Exadata Database Service、Oracle Exadata Cloud@Customer和Oracle Base Database Service的Oracle云基础设施远程软件服务 （Oracle Cloud Infrastructure, OCI)上, 以及Oracle Database@Azure上提供。 在Exadata上，已经支持向量数据之间的卸载（Offloading）。
By supporting Select AI, users can interact with the database in a natural language way, so that "Wensheng SQL" can realize queries, and the original factory support makes this function have a better experience in terms of performance and accuracy. Combined with the technical capabilities of Cohere and Llama, there are many more possibilities for OCI's generative AI applications, such as APEX, which makes SQL programming easier. With the help of converged databases, any form of data can be used in any environment and at any time to implement various business functions through the SQL capabilities derived from AI. In addition, customers can integrate Oracle Digital Assistant with industry applications such as HCM to meet specific scenario requirements.
In Oracle's words, Oracle Database 23ai is a "game-changing" release that will support up to a decade of service. "On the one hand, we continue to bring the latest technologies and products to the Chinese market, and on the other hand, we are also helping more and more Chinese companies to go overseas and integrate into the development of globalization." Wu Chengyang said, "All of this stems from Oracle's focus on enterprise-level IT technology, which is secure, stable, reliable, and advanced. This year marks Oracle's 35th anniversary in China, and we will continue to cultivate the Chinese market. ”
(8722530)