Today's guest post comes from Bob Thome, Vice President of Product Management at Oracle.
I recently read some interesting blog posts on the driving forces behind many of the today’s IT innovations. One of the common themes was the realization that sometimes purpose-built engineering is better at solving the toughest problems. Given 2018 marks the 10-year anniversary of the introduction of Oracle’s first engineered system, Oracle Exadata, I started thinking about many of the drivers that led to the development of this system in the first place. Perhaps not surprisingly, I realized Oracle introduced Exadata for the same reason driving other innovations--you can't reliably push the limits of technology using generalized "off-the-shelf" components.
Back in the mid-2000's, the conventional wisdom was that the best way to run mission critical databases was to use a best-of-breed approach, stitching together the best servers, operating systems, infrastructure software, and databases to build a hand-crafted solution to meet the most demanding application requirements. Every mission critical deployment was a challenge in those days, as we struggled to overcome hardware, firmware, and software incompatibilities in the various components in the stack. Beyond stability, we found it difficult to meet the needs of a new class of extreme workloads, that exceeded the performance envelopes of the various components. What we found was we were not realizing the true potential of the components, as we were limited by the traditional boundaries of dedicated compute servers, dumb storage, and general purpose networking.
Reviewing these objectives in light of the limits of the best-of-breed technology led to a simple solution--extend the engineering beyond the individual components and across the stack. In other words, engineer a purpose-built solution to provide extreme database services. In 2008, the result of this effort, Oracle Exadata, was launched.
The mid-2000’s saw explosive growth in compute power, as Intel continually launched new CPUs with greater and greater numbers of cores. But databases are I/O hungry beasts, and I/O was stuck in the slow lane. Organizations were deploying more and more applications on larger and larger SANs, connecting the servers to the storage with shared-bandwidth pipes that were fast becoming a bottleneck for any I/O intensive application. The economics and complexity of SANs made it difficult to provide databases the bandwidth they required, and the result was lots of compute power starved for data. The burning question of the day was, “how can we more effectively get data from the storage array to the compute server.”
The answer, in hindsight, was quite simple, although quite difficult to engineer. If you can’t bring the data to the compute, bring the compute to the data. The difficulty was you couldn’t do this with a commercial storage array—you needed a purpose built storage server that could cooperatively with the database process vast amounts of data, offloading processing to the storage servers and minimizing the demands on the storage network. From that insight, Exadata was born.
Over the years, we’ve built upon this engineered platform, refining the architecture of the system to improve performance, availability, security, manageability, and scalability, all while using the latest technology and components and minimizing overall system cost.
Customers have aggressively adopted Exadata, to host their most demanding and mission critical database workloads. Chances are you indirectly touch an Exadata every day—by visiting an ATM, buying groceries, reserving an airline ticket, paying a bill, or just browsing the internet. Four of the top five banks, telcos, and retailers run Exadata. Fidelity Investments moved to Exadata and improved reporting performance by 42x. Deutsche Bank shaved 20% off their database costs, while doubling performance. Starbucks leveraged Exadata’s sophisticated Hybrid Columnar Compression technology to analyze point-of-sale data while saving over 70% on storage requirements. Lastly, after adopting Exadata, Korea Electric Power processes load information from their power substations 100x faster allowing them to analyze load information in real time to ensure the lights stay on.
The funny thing about technology is you must keep innovating. Given today’s shift to the cloud, all the great stuff we’ve done for Exadata, could soon be irrelevant—or will it? The characteristics and technology of Exadata has been successful for a reason—that’s what it takes to run enterprise class applications! The cloud doesn’t change that. Just as in an on-premise world where people don’t run their mission critical business databases on virtual machines, because they can’t, customers migrating to the cloud will not magically be able to suddenly run those same mission critical business databases in VMs hosted in the cloud. They need a platform that meets their performance, availability, security, manageability and scalability requirements, at a reasonable cost.
Our customers have told us they want to migrate the Cloud, but they don’t want to forgo the benefits they realize running Exadata on-premises. For these customers, we now offer Exadata in the cloud. Customers get a dedicated Exadata system, with all the characteristics they’ve come to appreciate, but hosted in the cloud, with all the benefits of a cloud deployment: pay-as-you-go, simplified management, self-service, on-demand elasticity, paid for with a predictable operational expense budget with no customer-owned datacenter required.
However, not everyone is ready to move to the cloud. While the economics and elasticity are extremely attractive to many customers, we’ve repeatedly found customers unwilling to put their valuable data outside their firewalls. It may be because of regulatory issues, privacy issues, data center availability, or just plain conservative tendencies towards IT—they are not able or willing to move to the cloud. For these customers, we offer Exadata Cloud at Customer, an offering that puts the Exadata Cloud Service in your data center, offering cloud economics, with on-premises control.
So, it’s been a wild 10 years, and we are continuing to look for ways to innovate with Exadata. No matter whether you need an on-premises database, a cloud solution, or are looking to bridge the two worlds with Cloud at Customer, Exadata remains the premier choice for running databases. Look for continued innovation, as we adopt new fundamental technologies such as lower-cost flash storage and non-volatile memory, that promise to revolutionize the database landscape. Exadata will continue as our flagship database platform, leveraging these new technologies, and making their benefits available to you, regardless of where you want to run your databases.
I hope this post gives you a sense of the history behind Exadata, and some of the dramatic shifts that will be affecting your databases in the future. This is the first in a series of blog posts that will examine these technologies. Next, we will look more closely at performance, and why performance is critical in a database server, and how we’ve engineered Exadata to provide the best performance for all types of database workloads. Stay tuned for more:
Oracle Exadata Cloud at Customer: Bring the Cloud to your Business
Bob Thome is a Vice President at Oracle responsible for product management for Database Engineered Systems and Cloud Services, including Exadata, Exadata Cloud Service, Exadata Cloud at Customer, RAC on OCI-C, VM DB (RAC and SI) on OCI, and Oracle Database Appliance. He has over 30 years of experience working in the Information Technology industry. With experience in both hardware and software companies, he has managed databases, clusters, systems, and support services. He has been at Oracle for 20 years, where he has been responsible for high availability, information integration, clustering, and storage management technologies for the database. For the past several years, he has directed product management for Oracle Database Engineered Systems and related database cloud technologies, including Oracle Exadata, Oracle Exadata Cloud Service, Oracle Exadata Cloud at Customer, Oracle Database Appliance, and Oracle Database Cloud Service.