Oracle chairman and CTO, Larry Ellison, unveiled Oracle 18c, the world’s first and only fully autonomous database at the firm’s flagship OpenWorld event in San Francisco last night.
Oracle’s latest database is 100 per cent autonomous and automatically provisions, patches, updates and backs itself up without any human intervention. The technology is designed to automatically detect threats when they first occur then directs the database to remediate the problem immediately.
According to Ellison the significance of this is immense when considering the worst data thefts in history have taken place after a patch has been made available to prevent the cyber threat. By removing the need to wait for a human to schedule downtime — which involves taking a system offline — an enterprise’s defence against cyber threats is greatly augmented.
Moreover, Oracle’s latest database can patch itself immediately while running once it has been notified by the security system that a threat has been detected. This means no delay, and no human intervention is required.
The underlying technology powering Oracle 18c is machine learning — a technology Ellison described as every bit as revolutionary as the Internet itself and the first branch of artificial intelligence that actually works.
Ellison explained the firm’s advantage lies in the enormous amount of computer data it collects, including the billions of record and event logs produced every day. By studying all this information, Oracle’s systems can figure out what normal events look like and what an outlier looks like, by essentially unifying all this log data and training the computer to understand how to detect an anomaly.
“So if your CFO logs on at 3am in the morning in the Ukraine and you are a Californian retailer that is probably is not a good thing,” Ellison said.
“By training the computer to understand normal patterns from abnormal patterns we’re able to detect that it is highly abnormal that your CFO has logged on from the Ukraine with an IP address hidden by Tor.
“Tor’s whole purpose is concealing the intellectual property address and the real location it comes from but you detect any log on that goes through Tor.”
Ellison said information on normal patterns was equally as important as collecting information about anomalous events.
“By looking at normal patterns, we can automatically fine-tune the Oracle database, and we can tune it constantly. Data distributions could change, the size of the database could change, but we constantly adapt and fine-tune the system without human intervention.
“This is a big deal; no one else does this,” Ellison said.