SAN FRANCISCO--If everything Oracle CTO and co-founder Larry Ellison said the evening of Oct. 1 is true, then the company's board of directors, investors and stockholders had better have a meeting and find out whether Oracle will actually be able to make a profit from this new-fangled cloud-service business.
Ellison spent a good portion of his opening keynote at Oracle OpenWorld 2017 demonstrating how "cheap" Oracle's in-cloud workload processing is versus Amazon Web Services' RDS (relational database system). He explained in a series of demos that because Oracle's cloud service is anywhere from 6 to 15 times faster that AWS in processing the same workload, Oracle thus is 6 to 15 times "cheaper" than AWS.
Case in point: For the same "market research" workload, Ellison pitted an 8-node Oracle Autonomous Data Warehouse Cloud instance against a similar 8-node AWS DS2.xlarge cloud. The same eight queries were fed to both cloud services.
Timers were started. Oracle claimed AWS's processors took 244 seconds to do the job, costing the user 27 cents' worth of computing time. Oracle then claimed its own service took a mere 38 seconds to do the job, costing the user 2 cents' worth of cloudtime. This is going to be the new normal for the super-fast new DB, Ellison contended.
Oracle Says It Will Undercut AWS
Of course we know this example was a setup, but the CFO's question would be this: How in the world is Oracle going to make any money undercutting AWS--and, by inference, everybody else--by that much? There's a lot more to the story than this, for sure, but Oracle board members had better schedule an ad hoc meeting--and the sooner the better.
Oh, that's right: Ellison is chairman of the Oracle board. Maybe he won't call a special meeting after all.
Meanwhile on Oct. 1, Ellison introduced not only the aforementioned "world's first autonomous database cloud," but also an as-yet unnamed automated security product, which he said he would detail later in the week. He also used some of his time onstage to skewer Equifax following the security breach it suffered earlier this year in which more than 140 million people had their personal credit information compromised. We'll get to that in a minute.
Ellison said both the Autonomous Database Cloud and the security system use machine learning for automation to eliminate human labor, human error and manual tuning, and to enable availability, high performance and security at a much lower cost than competitors that include AWS.
“These systems are highly, highly automated--this means we do everything we can to avoid human intervention," Ellison said. "This is the most important thing we've done in a long, long time. The automation does everything.
Security is 'Robotic,' 'Autonomous'
"They're robotic; they're autonomous. In security, it's our computers versus their (hackers') computers. It's cyber warfare. We have to have a lot better computer systems, a lot more automation if we're going to defend our data."
Ellison said the automated security system would scan the entire system 24/7, know immediately when an intruder gets into it, and would be able to stop and isolate the intruder faster than any human can do it. He didn't mention that there are already systems out there that do the same thing, such as Vectra Networks, Vera and others.
On the DB side, Ellison said the database cloud eliminates human labor that touches tuning, patching, updating and maintaining the database. The company listed the following capabilities:
- Self-Driving: Provides continuous adaptive performance tuning based on machine learning. Automatically upgrades and patches itself while running. Automatically applies security updates while running to protect against cyberattacks.
- Self-Scaling: Instantly resizes compute and storage without downtime. Cost savings are multiplied because Oracle Autonomous Database Cloud consumes less compute and storage than Amazon, with lower manual administration costs.
- Self-Repairing: Provides automated protection from downtime. SLA guarantees 99.995 percent reliability and availability, which reduces costly planned and unplanned downtime to less than 30-minutes per year.
Oracle said the database cloud is designed to handle a high number of different workloads, including transactions, mixed workloads, data warehouses, graph analytics, departmental applications, document stores and IoT.
The first Autonomous Database Cloud offering, for data warehouse workloads, is planned to be available in calendar year 2017, Ellison said.
Details on Oracle's Autonomous Data Warehouse Cloud
Oracle Autonomous Data Warehouse Cloud ostensibly eliminates error-prone manual management tasks and frees up DBA resources, which can now be applied to implementing more strategic business projects.
Key features, according to Oracle, include:
- Simplicity: Unlike traditional cloud services with complex, manual configurations that require a database expert to specify data distribution keys and sort keys, build indexes, reorganize data or adjust compression, Oracle Autonomous Data Warehouse Cloud is a simple “load and go” service. Users specify tables, load data and then run their workloads in a matter of seconds—no manual tuning is needed.
- Performance: Unlike conventional cloud services, which use generic compute shapes for database cloud services, Oracle Autonomous Data Warehouse Cloud is built on the high-performance Oracle Exadata platform. Performance is further enhanced by fully-integrated machine learning algorithms which drive automatic caching, adaptive indexing and advanced compression.
- Elasticity: Oracle Autonomous Data Warehouse Cloud allocates new data warehouses of any size in seconds and scales compute and storage resources independently of one another with no downtime. Elasticity enables customers to pay for exactly the resources that the database workloads require as they grow and shrink.
Oracle Autonomous Database Cloud is powered by the company's latest database, Oracle Database 18c. Oracle Database 18c offers new automation capabilities in addition to enhanced OLTP, analytics and consolidation technologies.
Ellison on Equifax, Security
"You've got to know (about a breach) during the reconnaissance phase of a cyber attack," Ellison said, "when someone is nosing around in your computer system--trying to steal a password, trying to steal someone's identity. As they come in and start looking around, you'd better detect that that's happening."
Ellison chastised Equifax for not patching its system in time.
"I know it's a shock, but there was a patch available for Equifax, but somebody didn't apply it. I saw where the CEO lost his job--which doesn't bother me now, I'm not a CEO. That's a risky job those guys have," Ellison said with a slight laugh. "But no, I'd lose my job, too. It's a clean sweep (with a breach like Equifax's); directors aren't safe, nobody's safe when something like that happens."
This is going to get a lot worse before it gets better, Ellison said.
"People are going to get better at stealing data; we have to get better at protecting it," he said.
The Oracle Autonomous Cloud will become available on-premises or in the Oracle public or private clouds for data warehousing production workloads in December. It will become available for other specific workloads in June 2018.
For more information, go here.