Let’s talk about Oracle’s successful and expanding investment in cloud infrastructure. The company just celebrated its 45th anniversary, beat Wall Street’s estimated revenue in its fiscal fourth quarter, and showed its highest organic revenue growth rate in over a decade. The company is clearly doing a lot of things its customers like.
Front-and-center to Oracle’s success is Oracle Cloud Infrastructure (OCI) growth. Over the past year there has been a steady stream of OCI-related announcements. These have included plans to grow from 30 to 44 public cloud regions by the end of 2022 (39 are already in place), smaller Dedicated Region configurations, plans for Sovereign Clouds, new Cloud@Customer offerings, and expansions of OCI’s already impressive portfolio of services. This is perhaps the fastest expansion of cloud services by any service provider, and it helped drive Oracle’s 49% year-over-year IaaS growth and 108% growth in Exadata Cloud@Customer (Q4 FY22 earnings report).
And, if those aren’t enough to make you consider OCI for your public cloud, what about the new Oracle Database Service for Microsoft Azure that Larry Ellison and Satya Nadella announced at Microsoft Inspire on July 20th? This new service allows Azure customers to choose where to run Oracle Database for their Azure applications. Azure users can easily set up and use Oracle databases running on optimized OCI infrastructure directly from Azure, without logging into OCI.
The Oracle Database Service for Microsoft Azure is an Oracle-managed service currently available in 11 pairs of OCI and Azure regions worldwide. It uses the existing OCI-Azure Interconnect to offer latency between the two clouds of less than 2 milliseconds over secure, private, high-speed networks. This means that developers and mission-critical applications running on Azure can directly access the performance, availability, and automation advantages of Oracle Autonomous Database Service, Exadata Database Service, and Base Database Service running on OCI.
Oracle’s growth numbers represent a great metric to measure its overall success. However, most IT architects and developers want to understand why Oracle’s cloud offerings are better than the likes of Amazon Web Services (AWS) for their Oracle Database workloads.
The answer is simple. While Oracle is undoubtedly a strong competitor when matched head-to-head against nearly every public cloud offering, it offers clear advantages for Oracle Database applications. For example, organizations that use Oracle Database in their on-premises data center can more easily move workloads to OCI because it provides extreme levels of compatibility with on-premises installations and offers organizations the same or greater performance, scale, and availability. You won’t find a better example of this than Oracle’s cloud-enabled Exadata X9M platform that’s available natively in OCI or for Azure users through Oracle Database Service for Microsoft Azure.
Oracle Exadata X9M
Last year, Oracle delivered what may be the fastest OLTP database machine with the Exadata X9M. This machine is engineered to do only one thing: run Autonomous Database Service and Exadata Database Service faster and more efficiently than anything else on the market, delivering up to 87% more performance than the previous generation platform.
Wringing every ounce of performance and reliability from a database machine such as Oracle Exadata requires thinking about system architecture from the ground up. It requires a deep knowledge of Oracle Database and the ability to optimize the entire hardware and software stack. This is a job that only Oracle can realistically take on.
Exadata X9M’s employs a flexible blend of scale-up and scale-out capabilities that support virtually any workload by separately scaling database compute and storage capabilities. Of particular note is how the Exadata X9M provides high performance for both transactional and analytics workloads and efficient database consolidation.
Let’s start with analytics. At the highest level, Exadata X9M enables fast analytics through parallelism and smart storage. Complex queries are automatically broken down into components that are distributed across smart Exadata storage servers. The storage servers then run low-level SQL and machine learning operations against their local data, returning only results to the database servers. This allows applications to use 100s of gigabytes to terabytes per second of throughput—something you won’t find on your typical cloud database.
For OLTP, Exadata X9M breaks out some additional secret sauce in the form of scalable database server clusters, persistent memory (PMem) in the smart storage servers, and remote direct memory access over converged Ethernet (RoCE) that links them together. Databases run across hundreds of vCPUs to provide high performance and availability and read data directly from shared PMEM on the storage servers. The end result is that Oracle Database achieves SQL read latencies from shared storage of under 19 microseconds, which is more than ten times faster than traditional flash storage.
However, Exadata X9M in OCI doesn’t forego the use of flash memory, it embraces it. Without applications having to do anything, Exadata storage servers automatically move data between terabytes of PMem, tens or hundreds of terabytes of NVMe 4.0 flash, and terabytes to petabytes of disk storage to provide the best performance for different types of workloads. This results in a level of performance that isn’t possible with a traditional on-premises or cloud architecture built using generic servers and storage.
Bringing X9M to the Cloud
There’s no question that cloud resources are integral to nearly every enterprise’s IT infrastructure. The cloud offers a flexible and scalable consumption model with economics that can be superior to traditional on-premises deployments. While cloud infrastructure can be easily scaled to meet many growing application needs, this is not necessarily true for databases that support mission-critical applications. It’s common for organizations to have to refactor applications and redesign databases when they move to the cloud to provide the same levels of performance and availability they had premises, such as when moving Oracle Database to AWS. However, by deploying Exadata X9M in OCI, Oracle eliminates the expensive and time-consuming need to refactor applications for the cloud.
Oracle Exadata X9M in OCI shines for enterprise applications by delivering an elastic cloud database experience. For example, when running Autonomous Database Service or Exadata Database Service on dedicated X9M infrastructure in OCI, you can use 2 to 32 database servers and 3 to 64 smart storage servers in any combination. This means you can deploy platforms with more database servers for heavy OLTP workloads, more storage servers for data warehouses, or an even mixture of each when consolidating both types of workloads.
You can get the raw numbers for CPUs, storage, and memory for Exadata X9M in OCI from the Oracle website. Still, the critical thing to know is that all configurations deliver the database capabilities that enterprises require. For instance, the “entry” Exadata X9M configuration in OCI supports 19 microsecond SQL Read IO latency, 5.6 M SQL Read IOPS, and 135 GB/second of analytics throughput. Furthermore, with the ability to scale database servers by 16x and storage servers by 21x, we expect that no organizations will run into performance limitations.
Oracle tells us that by putting Exadata X9M into OCI, it now delivers the world’s fastest OLTP cloud database performance, and they have the data to back it up. Latency is critical for OLTP workloads, an area where the X9M has no equal. Exadata X9M’s 19 microsecond SQL IO latency is 25x better than when running Oracle Database on AWS Relational Database Service (RDS). The analytics throughput numbers from shared storage are even more impressive, with Oracle claiming that Exadata X9M in OCI delivers up to 384x the analytics throughput of Oracle Database running on AWS RDS.
Oracle has conquered the performance challenges for OLTP and analytics in the cloud and delivers this level of performance with attractive economics. Oracle makes the Exadata X9M for OCI available with a true consumption-based model where you only pay for the size of platform you need and the consumption you use. One key feature of Oracle Autonomous Database running on Exadata X9M is that it can auto-scale consumption by 3x based on the demands of the queries executing at every point in time. This helps you meet peak requirements by scaling up database consumption when needs grow and minimizes costs by scaling it back down later. Oracle cites global customers using these scaling capabilities to economically meet seasonal demands for retail companies and end-of-quarter financial closes for any business.
Running business workloads in the cloud is popular and continues growing at impressive rates because it solves practical problems for IT practitioners and business users. However, generic cloud infrastructure hasn’t delivered the same level of performance and availability for mission-critical OLTP and analytics workloads that many customers achieved with on-premises platforms.
If your enterprise depends on Oracle Database technology—and 97% of the Global Fortune 100 companies use Oracle Database, with 88% relying on Oracle Exadata for business-critical workloads—you need to seriously consider running your cloud database workloads on Exadata X9M in OCI. Oracle’s expanding portfolio of OCI services and delivery platforms, coupled with its unique ability to integrate optimized database platforms like Exadata X9M into OCI redefines what it means to run mission-critical databases in the cloud.
The Exadata X9M is built by the same people who build the Oracle Database, best positioning Oracle to optimize the performance, reliability, and automation required to get the most out of Oracle Database in the cloud. Oracle Exadata X9M is a stellar piece of engineering, bringing together compute and storage in an optimized architecture that delivers levels of throughput and reliability that deserve the superlatives I’m throwing around. And, it’s not just me saying it; Oracle’s momentum in the cloud bears this out as customers continue to make Exadata their preferred option to run Oracle Database.
When combined with the new Oracle Database Service for Microsoft Azure, Exadata X9M in OCI should cause organizations to rethink strategies focused on using generic cloud infrastructure for critical database applications.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.