OOW 2017: The self-driving database is here

For the past five OOWs at least, Oracle CTO, Larry Ellison, used to make a blizzard of product announcements, something which current CEO Mark Hurd himself had pointed out.

In essence, Oracle had been getting ready to provide end-to-end cloud services; all the way from apps to infrastructure to platform; and they had to build a lot of products to enable their cloud strategy – help customers to get on to the cloud, or to start in the cloud.

So, the product announcements made during OOW 2017’s Sunday keynote last night, was really a culmination of Oracle’s journey thus far.

During Day One of Oracle Open World 2017, it was all about databases and cybersecurity, including how Oracle is heading towards its ambitions to have these two fully automated.

A self-driving database is a big deal; think of the manual labour costs and human error risk which could be eliminated, not to mention the better agility and flexibility as a result. More details about Oracle’s Autonomous Cybersecurity product were revealed during Ellison’s second keynote on Tuesday.

But, during his first keynote, Ellison dubbed their self-driving database, the Oracle 18c, as the world’s first and only autonomous database.

Very, very awesome

Unreal: Ellison points out that a well-known cloud provider’s guarantee of reliability and availability is not real, and showed the provider’s list of exceptions in the fine print – downtime for planned maintenance, software bugs, downtime when adding compute and storage, downtime for upgrades and patching, and downtime for regional outages.

After the introduction, came a seemingly endless staccato of point after point, about why Oracle’s public cloud, with a standard configuration of Exadata X6 hardware, was better than Amazon Web Service (AWS’) own fully managed data warehouse, Redshift.

All in all, Ellison claimed that Oracle could guarantee 99.995-percent availability, with less than 30 minutes of planned and unplanned downtime, in a year.

Besides this, the CTO claimed that their new database (for data warehousing) which is due out this December, is very efficient, and instantaneously elastic. “You will never be provisioned more than you need.”  Being self-driven, this expanding or shrinking of resources would happen automatically, without downtime.

Ellison commented, “Cost can be minimised by delaying the resource allocation, until it is needed.”

This also leads to the notion that Oracle 18c does fully automated performance tuning, continuously. Yes, it does, and it will also do major upgrades while still running, Ellison had said.

Transparency? Oracle’s promise aka contractual guarantees,

“No one else has this stuff. We have been working on this for a very, very long time.”

In contrast, Redshift can’t automatically increase resources to run bigger workloads, and then later free them up. He said, that it is not an elastic system as they claim, and that they would have to shut down the system, just to be able to add resources.

Depending on whose workload is on whose database, AWS overall costs 9 to 16 times more than Oracle.

BUT, this does not include the rate card pricing for the underlying hardware, and if it’s Oracle’s highest-end, engineered system, Exadata, the pricing for that should be taken into consideration against Amazon’s hardware pricing.

Nevertheless, during Ellison’s demo, Oracle’s Exadata-based performance creamed the competition.

The biggest burn of all, came when Ellison said, “It’s not unusual for our competitors to use our technology. Amazon knows this. They are one of the biggest Oracle users on the planet Earth. SAP is one of the biggest users of Oracle on Earth.”


(This journalist is a guest of Oracle’s to their annual conference in San Francisco).




Oracle Debuts Revolutionary New Machine Learning Applications in Opening Keynote at Oracle OpenWorld 2017

There are no comments

Add yours