More powerful server hardware is coming to the Exadata database service, bringing increased scalability and throughput for Oracle's database and analytics workloads.
Oracle released the latest iteration of its Exadata Cloud Service today, providing users with increased database scale and performance.
The updated Exadata Cloud Service X8M is set to be generally available later this month, the vendor said.
Exadata Database Machine is Oracle's purpose-built database hardware server platform, while the Exadata Cloud Service makes Exadata systems available on a consumption basis in the Oracle cloud.
The Redwood City, Calif., tech giant first introduced the Exadata X8M hardware in September 2019. Now Oracle has rolled out X8M across 26 global cloud regions on the Oracle Cloud Infrastructure platform.
The Exadata X8M hardware uses Intel Optane persistent memory to enable more data to reside in-memory, providing faster access to data. The X8M also gets advanced connectivity capabilities including 100 gigabit remote direct memory access (RDMA) over Converged Ethernet. With the updated Exadata Cloud Service, the X8Ms are helping to enable a new level of scalability for database workloads in the cloud. According to Oracle, a database workload running on Exadata Cloud Service X8M can have up to 25 petabytes of capacity spanning up to 4,600 CPU computing cores.
Accelerating workloads with Exadata Cloud Service
"Enterprises know they need to run more efficiently, practicing enterprise acceleration," said Holger Mueller, an analyst at Constellation Research. "Their databases are a key aspect in this, and being able to run the Oracle database more efficiently both from a technology and commercial standpoint is a big win and that's what Oracle delivers with Exadata X8M cloud service."
Oracle's Exadata Cloud Service enables users to provision Exadata database machines as needed to handle workloads.
While there are other cloud database services, Mueller said he doesn't see anything directly comparable to what Oracle offers. He noted that if Microsoft or IBM would provide custom hardware for Microsoft SQL Server or IBM Db2, that would be an alternative, but that doesn't exist.
"Oracle, knowing its own software better than anybody else, can build the perfect hardware for its database," Mueller said.
Enterprises know they need to run more efficiently, practicing enterprise acceleration. Their databases are a key aspect in this, and being able to run the Oracle database more efficiently both from a technology and commercial standpoint is a big win.
Holger MuellerAnalyst, Constellation Research
Exadata Cloud Service versus Autonomous Database
Juan Loaiza, executive vice president of mission-critical database technologies at Oracle, noted during a press and analyst virtual briefing that Exadata is the hardware platform that Oracle has been working on for over a decade, to make it an optimized platform for database workload.
The Oracle Exadata Cloud Service is basically the Oracle database running on the Exadata hardware, but with all the attributes of cloud, such as elastic scaling for demand and consumption-based pricing at a per-second-of-usage rate, Loaiza said.
Oracle Exadata Cloud Service is not the same as the Oracle Autonomous Database cloud service. The difference between the two cloud database services, according to Loaiza, is what Oracle refers to as user-controlled database automation.
"With autonomous database, the users kind of sit in the backseat and we take care of everything," he said. "With the Exadata cloud service we provide the automation, but the user is still in control."
X8M hardware boosts cloud database performance
From a performance perspective, Loaiza noted that with the X8M based Exadata Cloud Service, users will get 100 gigabit per second converged Ethernet fabric, a jump up from the 40 gigabits in the prior version.
"That's two and a half times higher throughput and it's really good for analytics," he said.
Also, Loaiza said that the RDMA technology combined with persistent memory further accelerates workloads. The database servers directly access the persistent memory in the servers through RDMA, bypassing all the software layers. The benefit is increased transaction processing performance.
With the cloud service, Loaiza noted that as users need to scale up workloads, new systems can be added to a database cluster. As the cluster scales up, the workload will be rebalanced across all the systems.
"That's one of the big pluses of a public cloud. We have all this hardware available and anytime you want, you can just kind of add to it," he said.
_mobile.jpg)
