Oracle chairman and CTO Larry Ellison last week said he’d do it, and he did.
The company best known for its conventionally licensed, on-premises database is launching itself as a faster, cheaper competitor in the infrastructure-as-a-service arena, going up directly against market leader Amazon Web Services and the other three principal contenders in the cloud: Microsoft, Google and IBM.
And, in typical Ellison style, the tech pioneer is not mincing words about the top dog in cloud computing, declaring that Amazon’s “lead is over” and that they are about to get some “serious competition going forward,” according to a recap by ZDNet and USA Today of Ellison’s talk Sunday during a keynote address at the OpenWorld conference in San Francisco.
“Amazon was the innovator, the first mover, and I have a lot of respect for them, but now we’re aggressively moving into infrastructure,” Ellison said Sunday. “Our new data centers offer twice as many cores as Amazon, twice as much memory as Amazon, four times as much storage as Amazon, and more than ten times the I/O capacity of Amazon. But you have to be willing to pay less.”
Ellison is no stranger to taking aim at big competitors, and the technology executive long battled and publicly railed against Microsoft. Now, it appears that Ellison is taking some of that competitive spirit, and directing it toward Amazon.
Though Oracle’s spate of announcements focus on IaaS, their thrust is helping organizations to set up a hybrid environment, splitting computing between their own data center and the cloud and more easily moving data and apps between those locales. Ellison predicted a period of at least 10 years during which on-premises and cloud computing will co-exist for most organizations.
One of Oracle’s new offerings — dubbed Bare Metal Cloud Services — includes generic servers that the company claims are 11.5 times faster and 20 percent cheaper than the fastest offerings from competitors. Bare Metal Cloud Services offers database-as-service, network block storage, object storage and support for virtual private networks. It will be available in three independent geographical areas connected by a fiber-optic ring, with automatic failover among them, Ellison said.
Services will be available in North America and Europe.
Oracle also announced its new Database-as-a-Service (DBaaS) offering, based on Oracle Database 12c Release 2. The new release will be available as a cloud service first, Oracle stressed.
The company previewed what it’s calling the Ravello cloud service: a container service claimed to allow running VMware and kernel-based virtual-machine workloads in the public cloud, with no need to convert virtual machines, reconfigure the apps or play with network settings.
Oracle also announced FastConnect, a way for customers to connect their data centers to Oracle’s cloud. It offers IPsec virtual private networking, MPLS to connect to existing enterprise networks and Fast Connect for quick peering, Oracle said.
Despite Ellison’s bluster, Oracle has a long way to go before posing a threat to AWS, Azure or even Google Cloud Platform.
It’s true that for the quarter ended August 31, delivering Oracle’s DBaaS (plus platform as a service) together yielded $798 million in revenue (9 percent of total revenue), up 77 percent year over year. That’s strong evidence of the trend toward using databases more in the cloud and less on-premises. And IaaS revenue was $171 million (2 percent of total revenue), up 7 percent.
But at the moment, Oracle barely figures into Gartner, Forrester and other analysts’ assessments of the IaaS market.
By far the largest chunk of its revenue (68 percent) last quarter was still coming from its on-premises software. AWS strongly leads the IaaS field, followed by Azure. Oracle lacked the market share to even figure into Gartner’s most recent IaaS magic quadrant analysis.
As Azure executive VP Scott Guthrie remarked last week at a Deutsche Bank technology conference, the time to enter into real IaaS may be through.
“To be a hyperscale cloud vendor, you need three things,” he said. “You need to be able to spend vast sums of money, in the billions; you need thousands of engineers to write distributed-systems code; and you need time. The combination of those things creates a pretty big moat, and I think it’s going to be very difficult for folks to break in.”