Oracle Corp. Sunday announced the next generation of its database, which founder and Chief Technology Officer Larry Ellison said will be able to handle key tasks such as critical software patches automatically.

Ellison (pictured) also said that the autonomous database will help on another product to be introduced in detail Tuesday: new cybersecurity technology that will offer automated threat detection and immediate fixes for the threat. Both the new database and the new cybersecurity services are driven by machine learning, the branch of artificial intelligence that allows computers to learn on their own.

Ellison made the announcements in a Sunday evening keynote to open Oracle’s annual OpenWorld conference this week in San Francisco, where he took yet another opportunity to boast that Oracle is ahead of cloud computing leader Amazon.com Inc. “It’s the most important thing we’ve done in a long, long time,” he said.

The “self-driving database,” as Ellison likes to call it, will handle tasks such as updates and software patches automatically to reduce the high cost of human labor, he said. The new database, called the 18c Autonomous Database, will be able to patch itself while running, he said, without having to wait for a human being to find a time to do a patch. It also will be able to be provisioned, do upgrades and tune itself while running, he said.

Ellison also promised that it will cost half as much to run Oracle’s database as it does on Amazon Web Services Inc.’s cloud. The new database will be available starting in December for data warehousing, with the version for online transaction processing planned for release next June. Both will be available in Oracle’s public cloud and on-premises.

In additional, Ellison introduced a new pricing model for the new Autonomous Database that he called “bring your own license to PaaS,” which allows users to get the new database up and running for $300 per month with a minimum configuration of one OCPU, which is the equivalent of an Intel Corp. Xeon processor, and 1 terabyte of storage. He added that this pricing model applies not only to the database, but also to Oracle’s middleware, PaaS and analytics services.

Oracle’s new database is more “elastic” than Amazon’s rival Redshift database, he said, meaning it can adapt rapidly to workloads without wasting resources in the process. “Amazon’s database, Redshift, cannot automatically increase the number of processors to run a bigger workload and then free up those processors. It just can’t do it,” Ellison said.

In a benchmark test for an insurance analysis workload using the new model, Ellison compared the cost and performance of the 18c Autonomous Database to Redshift. Not only did Oracle complete the workload faster – 26 seconds versus 282 seconds for Amazon – it also did so much more cheaply, with Amazon’s service proving to be 8.3 times more expensive.

“If you want jobs to be done faster, you’ve got to be willing to pay less,” Ellison joked.

Ellison also boasted of Oracle’s much higher availability with its new database, claiming 99.995 percent uptime over the course of a year, amounting to just 30 minutes of downtime annually. Amazon makes similar claims, of course, but Ellison claimed Amazon’s guarantee “is not real” due to the vast number of exclusions detailed in its contracts. “[Amazon’s] guarantee doesn’t count when you’re down for any reason,” he said. “It doesn’t include downtime or maintenance, or patching, and certainly doesn’t cover downtime if there’s a bug.”

Not surprisingly, an AWS spokesperson begged to differ. “Yeah, that’s factually incorrect. With Amazon Redshift, customers can resize their clusters whenever they want, or can scale compute separately from storage by using Redshift Spectrum against their data in Amazon Simple Storage Service and pay per query for just the queries they run,” AWS said in a statement. “But,‎ most people know already that this sounds like Larry being Larry. No facts, wild claims and lots of bluster.”

The autonomous nature of Oracle’s new database means there are additional cost savings that are more difficult to quantify, Ellison said, pointing to the reduced workload for database administrators. “You’re automating away the job of database professionals,” he said. “So we’ll see a migration of database skills to focus more on schema design, analytics and setting policies for what’s mission-critical. Database pros will have a lot more time to focus on securing data.”

In broad outline, the main announcement of an autonomous database was no surprise. Ellison spilled the beans during Oracle’s earnings report in August, and again last month while announcing new cloud pricing at a customer event. And machine learning is hardly a new development, having led in recent years to everything from instant speech translation to self-driving cars.

Since before the last Open World in 2016, Oracle had gone on the offensive against Amazon, most recently on Sept. 19 with a new cloud pricing plan that gives discounts to Oracle database customers who move their databases to the cloud. The autonomous database is the second part of a plan to make Oracle’s database run for 50 percent less cost on Oracle’s cloud than on AWS’s.

Not everyone, of course, buys Ellison’s line. “Larry’s self-driving database completely misses the point,” said Harry Glaser, co-founder and chief executive of Periscope Data, a top Amazon Redshift customer that helps data scientists created detailed visualizations of data. “Businesses don’t need help automating database administration. What they really need is better software that helps data teams – from data analysts and engineers to data scientists – do their jobs better.”

Ellison’s focus on machine learning, already in use for years by many companies, also might be a little opportunistic. “A lot of companies are trying to grab on to a little bit of the machine learning/artificial intelligence sparkle dust,” said Doug Henschen, vice president and principal analyst at Constellation Research.

For all the news on its database, Oracle’s future depends most of all on moving its software and its customers to its cloud. Oracle remains far behind leaders such as AWS and Microsoft Azure and even other upstarts such as Google Cloud Platform, at least in the base-level infrastructure as a service that provides computing and storage services. “They have to succeed in the cloud,” said Henschen.

That’s one reason the company has been focusing on other layers of the cloud computing “stack,” such as platform as a service and software-as-a-service, or online business applications, such as Salesforce.com Inc. and Workday Inc. provide. Oracle’s database is certainly an established PaaS service, and it also offers a wide range of enterprise resource planning, supply chain management and other core business software as a service.

Oracle is also hoping to turn its continuing strength inside data centers into an advantage by pitching itself as the best choice for the majority of companies that need to do computing not just in the cloud but in their own data centers as well because of the cost of moving data or applications or regulatory limitations on where data resides.

The company says it’s the only one that can provide soup-to-nuts technology to manage that so-called hybrid cloud setup. It offers hardware appliances to mirror cloud services inside data centers as well as software and cloud services, promising the services will work the same on-premises or off.

Photo: Oracle live stream