Oracle Exadata X10M processor offers major performance leap

Oracle Exadata X10M processor offers major performance leap

Oracle today unveiled the latest generation of Exadata processors. It promises with X10M to deliver big performance improvements with AMD’s 4th generation EPYC chips, paying off the Intel Xeons of the X9M.

Below the line, Oracle has seen transaction throughput improve by a factor of three. In addition, analytic query performance is 3.6 times better than before. In the real world, this means customers can deliver faster services. For example, many financial institutions use Exadata processors to process transactions. In addition, analytic query is vital for getting results from one or more data sources, for example, to assess international sales data.

Less space, less money

The inclusion of the Zen 4-based EPYC chip allows Oracle to utilize 96 cores in a single socket. In addition, the associated storage server has 22% more capacity than before. All-flash storage has gone up 2.4 times. In addition, database servers now have 50 percent more memory available, allowing more databases to run on a single system. The platform supports up to 3 TB of DDR5 memory. The number of network interface cards Oracle supports has gone from three to five.

All this means Exadata X10M take up less space than X9M required. In addition, the pricing model is consumption-oriented, so customers benefit from the efficiency gains.


The platform is optimized on all fronts for Oracle database workloads, states Ashish Ray, Exadata’s VP of product management, Oracle Autonomous Database, database cloud and high-availability computing.

Where Ray describes the speed improvements more precisely to SiliconANGLE, he touts a network card called Remote Direct Memory Access Over Converged Internet. “When the database servers need to access the storage servers, they are accessing storage server memory as if it’s memory mapped to input/output. They are not going through the standard TCP/IP connection or doing a CPU interrupt or context switch. The latency can be on the order of microseconds.”

As for the choice customers have, they can choose independent numbers of storage and database servers. Ray indicates that users can therefore choose something based on their specific workloads, so more storage for I/O-intensive tasks and more database servers if it is compute-intensive.


Oracle is available both on-prem with Exadata Cloud@Customer as well as in the cloud through Oracle Cloud Infrastructure. Both involve the same hardware and customers can control them in the same way.

Regardless, Oracle’s cloud plans have continued to expand. For example, with EU Sovereign Cloud, it is going to offer a solution that meets GDPR requirements on our continent. In addition, we recently spoke with Richard Smith, vice president of technology EMEA at Oracle, about the company’s strategy on generative AI.

Tip: Oracle’s strategy for generative AI revolves around infrastructure and apps