Oracle, a leader in cloud technology, has introduced the Oracle Exadata X11M, the newest version of its Exadata platform. According to the company, this model offers up to 55% quicker AI vector searches, 2.2 times faster analytics scan throughput, and 25% improved transaction processing, all while maintaining the same pricing as its predecessor.
“Exadata X11M is up to 70 times faster due to the unique hardware offloaded RDMA compared to AWS RDS and Azure SQL. Now, with multi-cloud, we’re faster than AWS in AWS and faster than Azure in Azure,” said Kothanda Umamageswaran, Oracle SVP of Exadata and scale-out technologies. He explained that Exadata offers lower latency for OLTP database operations, handling tasks faster than AWS RDS and Azure SQL, which take around 1 millisecond compared to Exadata’s mere 14 microseconds.
Oracle’s multi-cloud strategy allows them to deploy their Exadata technology on competing cloud platforms, claiming superior performance even in those environments, including on AWS and Azure infrastructures. It challenges these providers in their own ecosystems.
Currently, Oracle Exadata X11M is available on-premises, with Exadata Database Service, Autonomous Database on Exadata Cloud@Customer, Oracle Cloud Infrastructure (OCI), and in multi-cloud setups including AWS, Google Cloud, and Microsoft Azure.
In an exclusive interview with AIM, Ashish Ray, Oracle’s VP of mission-critical database technologies, revealed that their customer base largely consists of on-premises clients, followed by those using Oracle Cloud, with multi-cloud customers ranking next.
“The Google and AWS announcements are fairly recent, so naturally, this multi-cloud play is not at the same scale as our current on-premises and Oracle Cloud offerings,” Ray said. However, the company anticipates significant growth in this multi-cloud initiative, including Azure, with the hope that clients using other platforms will move to its high-performance database.
“The walled gardens have come down, and business productivity has become much more transparent. There are no bottlenecks, no setbacks, and now the choice is back to the customers,” said Ray.
He described India as a data-driven market experiencing explosive growth, fuelled by the digital modernisation of businesses—from large enterprises to a vibrant startup ecosystem. Ray expects that in India, Exadata X11 will act as a catalyst for further business innovation and IT modernisation.
Furthermore, he added that migrating to Exadata X11M provides better value for businesses. “With Exadata X11M, companies can handle more workloads with fewer servers, which helps reduce costs related to data centres, power usage, space, and cooling, especially for those using on-premises setups.”
Meanwhile, Umamageswaran notes that companies requiring access to vast amounts of data from multiple sources and data types, say, graph, spatial, documents, transactional, and analytical data, will benefit from Oracle Databases 23ai on Exadata X11M.
The company claims that Exadata handles transactions in microseconds and also supports real-time analytics like anti-money laundering (AML) checks that must occur simultaneously with the transaction process.
Database Competition Galore
The catch with Oracle Exadata is that it is only compatible with Oracle databases such as Oracle Database 23ai. But today, customers are experimenting with different database platforms such as Amazon Aurora, Cosmos DB, Google Spanner, MongoDB, CockroachDB, and many more.
“There will always be individual data management solutions, whether it’s PostgreSQL, Aurora, Redshift, or Cosmos DB,” Ray added.
Much like Oracle, CockroachDB is a distributed SQL database designed for scalability, consistency, and resilience. It combines the best features of traditional relational databases with the advantages of modern distributed systems.
Spencer Kimball, CEO of Cockroach Labs, in the latest episode of Tech Talks at AIM, critiqued the dependence on cloud vendors like AWS, Google Cloud, and Oracle for databases, emphasising the limitations of vendor lock-in. He stated:
“With CockroachDB, we let customers run across clouds, in private data centers, or both. You can even turn off a cloud vendor entirely and keep running without downtime. That’s true freedom, and it’s what enterprises need today.”
To this, Ray said: “A lot of these companies tend to be SQL-compatible. We do have customers who run CockroachDB and PostgreSQL,” adding that the scale at which Oracle Database runs and supports large enterprises across the entire globe is absolutely unheard of. “That’s why you see a lot of multi-cloud interest from the cloud hyperscalers,” he added.
From a generative AI perspective, he highlighted that Exadata significantly accelerates tasks like vector processing, index creation, and querying.
However, he did acknowledge the growing popularity of PostgreSQL due to its open-source nature. Oracle Database doesn’t directly support PostgreSQL, but Oracle offers a service called OCI Database with PostgreSQL. This service lets users use PostgreSQL in the Oracle Cloud, giving them PostgreSQL features along with Oracle’s cloud benefits.
AMD is EPYC: Oracle launched Exadata in 2008. Ray explained that while the Oracle Exadata X9M was based on Intel processors, the new Exadata X11M uses AMD EPYC processors. “The reason for moving to AMD was that it was just available with numbers that can drive a lot of parallel throughput for database workloads,” said Ray.
Interestingly, NVIDIA also recently partnered with AMD. AMD’s EPYC CPUs are critical in powering NVIDIA’s GPUs for large-scale AI workloads. “We’ve shown a 20% improvement in training and a 15% improvement in inference when connecting EPYC CPUs to NVIDIA’s H100 GPUs,” said Ravi Kuppuswamy, senior vice president & general manager at AMD.
