Oracle Exadata update boosts performance to meet AI needs
With database workloads growing due to the demands of AI development and real-time analytics, the tech giant's latest database infrastructure update focuses on efficiency.
The latest version of Oracle's Exadata platform is now generally available, adding performance boosts for AI, analytics and online transaction processing workloads on the tech giant's relational database.
Exadata, first introduced more than a decade ago, is the infrastructure for Oracle Database and has been updated numerous times to remain current as data management, analytics and AI have evolved.
Oracle Exadata X11M, which was launched Tuesday and is available at the same consumption-based pricing as previous Exadata versions, targets better efficiency. The update combines AI to manage usage with the ability to run workloads faster and on fewer systems, which results in better performance, according to Oracle.
Beyond efficiency, Oracle's Exadata update aims to provide flexibility by making the tech giant's relational database capabilities available in various environments, including public cloud, multi-cloud and on-premises.
Given the greater efficiency provided by the update while keeping prices the same, it is a significant advance for the tech giant's customers, according to Holger Mueller, an analyst at Constellation Research.
"Oracle keeps passing the Moore's Law dividend to customers of better performance at the same price," he said, referencing the notion posited by Intel founder Gordon Moore that as workloads increase so will efficiency. "It is becoming an Exadata [recurring theme] across releases."
With that continued focus on performance, Oracle Database is outpacing rival databases such as IBM Db2 and Microsoft SQL Server, Mueller continued.
"The Oracle Database on Exadata has won," he said. Oracle provides an array of data management and analytics platforms, including database offerings Oracle Database and Autonomous Database.
In July 2024, Oracle unveiled Exadata Exascale, a new architecture for Oracle Database designed to meet the needs of enterprise customers as their increasing investment in AI development places new demands on the performance of databases and other data management tools.
The update
Given its potential to enable widespread use of analytics, which for decades was limited to trained experts, and exponentially improved efficiency, enterprises are investing heavily in generative AI applications.
However, both generative AI as well as traditional AI require far more training data to be accurate than an analytics tool, such as a dashboard or report. Without copious amounts of data, generative AI applications are more likely to hallucinate than those trained with the right amount of data.
As a result, the workloads needed to train AI applications are bigger than those for other data-driven applications. To meet the workload requirements of AI development without slowdowns, many database vendors have made performance improvements a priority.
For example, Neo4j targeted performance speed in late 2023 and InfluxData did the same in late 2024.
Oracle has similarly prioritized performance improvements, with Exadata X11M the latest example. By providing the improved version of Exadata without raising prices, the tech giant is not only serving its existing customers but also could attract new ones, according to Shawn Rogers, an analyst at BARC U.S.
Speed and flexibility … are valuable enhancements with logical return on investment and cost savings. Starting at the same price as Exadata X10M makes it a no-brainer for many Oracle customers and prospects, especially those interested in powering AI and analytic applications.
Shawn RogersAnalyst, BARC U.S.
"Speed and flexibility … are valuable enhancements with logical return on investment and cost savings," he said. "Starting at the same price as Exadata X10M makes it a no-brainer for many Oracle customers and prospects, especially those interested in powering AI and analytic applications."
Specifically, Oracle's Exadata update adds the following, according to the vendor:
Up to 55% faster vector index searches and 43% faster in-memory vector index queries compared with Exadata X10M, enabling users to speed the discovery of the relevant data needed to train a model or application and lower usage-based pricing expenses.
Up to 25% faster serial transaction processing and 25% greater concurrent transaction throughput, which also leads to lower costs by enabling users to run workloads with less compute power.
Up to 25% faster analytics query processing, enabling users to run real-time analytics on transactional data.
Among the upgrades, substantially faster vector search is significant, according to Rogers.
Vector search, which enables the discovery of relevant unstructured data such as text and images, has become a critical part of AI development. More efficient vector searches, therefore, stand to benefit enterprises building AI models and applications.
"AI is driving innovation across the enterprise landscape," Rogers said. "Increasing vector search speeds … will enhance performance for Oracle customers, making AI applications more valuable and decreasing costs. Cost is the No. 1 challenge for companies leading in AI adoption per BARC research."
Adding deployment flexibility also is important, as enterprises have different requirements for different types of data, he continued.
Mueller also highlighted the improved vector search capabilities, given its prominent role in AI development.
"It is critical for Oracle to make sure the Oracle Database is relevant in AI [development], and Exadata is its platform for AI workloads," he said.
Regarding the reason Oracle made performance the focal point of its latest Exadata update, many customers are collecting more data than ever before and using their databases for more workloads, according to Ashish Ray, the tech giant's vice president of products for its database technologies, including Exadata.
"To effectively handle this evolving and complex data landscape, our customers need extreme levels of performance, scale, availability and security from their database platform, as well as the ability to control costs, meet sustainability goals and deploy solutions in diverse locations," Ray said.
Plans
With Exadata X11M now available, Oracle will focus on further performance improvements as database workload demands continue to increase along with growing data volume and rising investments in AI development, according to Ray.
Mueller said Oracle would be wise to develop a real-time integration between its databases and Oracle Intelligent Data Lake, which was unveiled in September.
Rogers, meanwhile, said that Oracle has positioned itself well relative to AWS, Google Cloud and Microsoft. Continuing to focus on performance speed and efficiency, cost savings and enabling AI and analytics application development would be wise.
"Oracle should -- and likely will -- double down on these priorities moving forward," Rogers said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.
