Oracle offers price-performance boost with Exadata X11M update

The new version of Oracle’s Exadata platform, X11M, is now generally available, and promises improved performance on generative AI tasks such as retrieval augmented generation (RAG) for the same price as the old version.

Exadata is Oracle’s hardware and software platform for running Oracle Database workloads on premises or in hybrid- and multi-cloud environments. The previous version, X10M, was released in June 2023, just as demand for generative AI was taking off, and some of the changes in the latest version address the needs of that market.

“The improvements inside X11M that support handling vector indexes faster at scale will help enterprises take on more demanding and complex generative AI use cases, such as advanced RAG and inferencing workloads that involve multi-step semantic search, adversarial guardrails, and agentic processes with less hardware than before,” said Bradley Shimmin, chief analyst at Omdia.

“This release really focuses on workload consolidation and optimization, something that will resonate well with companies looking to lower operational costs while adopting generative AI,” he said.

What is new in Exadata X11M?

Exadata X11M includes software and hardware optimizations to deliver faster vector search for AI, faster input output operations per second (IOPS) resulting in lower latencies for transaction processing, and faster data scans than X10M, said Oracle Vice-President of Product Ashish Ray, adding that the prices of both generations are exactly the same.

The company said that X11M offers up to 55% faster persistent vector index searches and 43% faster in-memory index queries.

For online transaction processing (OLTP) workloads, the new version offers 25% faster serial transaction processing and greater concurrent throughput resulting in 21% lower read latency, Ray said, and also processes analytic queries 25% faster.

Energy efficiency

With the increasing power demands of generative AI applications in the spotlight, Oracle also said that X11M helps improve the energy efficiency of IT operations as its superior performance compared to previous generations means fewer systems are needed.

“This will help enterprises save on infrastructure, power, cooling solutions, and data center space,” Ray explained.

Other software optimizations in X11M will allow enterprises to consolidate their workloads on smaller systems and manage power more efficiently, the company said.

“Intelligent power management is built directly into Exadata X11M, helping enterprises address energy efficiency and sustainability goals by letting them turn off unneeded CPU cores, cap power consumption, and optimize power utilization during periods of low usage,” Ray explained.

What do these upgrades mean for enterprises?

The upgrades in Exadata X11M will enable enterprises to take on projects that were not possible previously, owing to latency or cost factors, Shimmin said.

Another utility of X11M, according to Moor Insights and Strategy principal analyst Matt Kimball, is the promise of helping enterprises turn raw data into value faster, which in turn can lead to an increase in revenue.

“Imagine if an enterprise can allow an end customer to take a decision 25% faster. Or a researcher being able to predict or forecast faster in their respective fields. Think about real-time trading desks and what a 25% improvement in performance can mean in terms of profitability,” Kimball said. 

The superior performance and lower operational costs of X11M will drive initial adoption across financial, e-commerce, and process manufacturing segments where demand to scale and performance across complex workloads, such as Oracle, SAP, and Siebel, are key, said The Futurum Group’s principal analyst Ron Westfall.

However, dbInsight’s chief analyst Tony Baer pointed out that demand for AI is spreading across all sectors, and other industries such as healthcare could also look at adopting the new hardware-based platform.

Kimball suggested that Oracle might be using X11M to defend its dominance in the market against the likes of Teradata, AWS, and Microsoft.

Available across hybrid and multi-clouds

Oracle said that Exadata X11M can be deployed on-premises, or with Exadata Database Service and Autonomous Database on Exadata Cloud@Customer, Oracle Cloud Infrastructure (OCI), and multi-cloud environments.

This move, according to analysts such as Kimball and Westfall, is aimed at helping its enterprise customers seamlessly integrate their application and data environments.

Moreover, Westfall said thar the combination of on-premises and cloud support aligns with enterprise customer demand to meet their data sovereignty and compliance requirements that includes keeping sensitive data on-premises alongside leveraging cloud services.

The support for hybrid cloud deployments also enables customers to consolidate data centers while minimizing architectural shifts, decreasing costs while improving resource optimization, the analysts said.

In addition, Oracle said that enterprises that are running Exadata workloads currently can move to different deployment environments without any downtime.