Oracle’s executive chairman Larry Ellison spent a good deal of time talking about his long-time competitor IBM during his keynote speeches at his company’s tech conference this week in San Francisco.
That’s not unusual. Ellison loves a bit of trash talking and usually throws an entertaining barb or two at his rivals.
But this time, he’s been almost nostalgic about IBM, while also depicting IBM as a nobody when it comes to cloud computing. This is mostly on the basis of how many often Oracle salespeople compete against IBM’s for a customer contract.
During both of his keynote speeches this week, he offered this backhanded complement:
“IBM, arguably the greatest company in the history of companies — we never see them in the cloud. They were our major database competitor. They were our major middleware competitor. We never see them. This is a completely different world and a next-generation of technology.” (Middleware is the software used to run and manage other software applications.)
But there’s good reason why Oracle rarely competes head to head with IBM on cloud contracts these days: they are really in very different cloud markets.
Oracle is bringing its enterprise apps to the cloud
Oracle is mostly focused on taking its massive catalogue of existing enterprise applications, re-writing them for the cloud, and making them available to rent as a service. This is known as software-as-a-service (SaaS). This is also where Oracle is having the most cloud success.
To really do SaaS right, Oracle needed to offer a service that lets app developers create apps in the cloud, known as platform-as-a-service (PaaS), Ellison said. Oracle controls one of the most popular web development languages, Java, and so it has its own network of about 9 million developers to tap for its PaaS service.
To really do PaaS right, it needed a service that let’s companies and developers just rent compute power, storage, memory and operating systems and use that stuff as they wish, known as Infrastructure-as-a-Service (IaaS).
The biggest player in IaaS is Amazon, with Microsoft coming on strong.
Just about anytime Oracle is competing for an IaaS contact, it tends to run into the 800-pound gorilla Amazon. Nearly every enterprise would put Amazon on its short list.
IBM wants ‘cognitive computing’
IBM is also a somebody in IaaS, thanks to its purchase of Softlayer a few years back. But IBM CEO Ginni Rometty is focusing her cloud efforts on an entirely different vision than Oracle. Instead of trying to win enterprise’s SaaS application business, or their IaaS business, she’s trying to create a new thing in the cloud that she calls “cognitive computing.”
Cognitive computing is like turning your computer into a human-like advisor. You take lots of your own company’s data, mix in data from other sources (like Twitter, weather data, etc.) and put all that into a supercomputer that can think, learn, talk and reason like a human. That’s a tech called machine learning. And IBM Watson is arguably the most popular form of cloud computing machine learning available today.
Rometty is trying to build a completely new computing services to make predictions and offer advice.
Instead of just keeping track of your transactions in a database in the cloud, Rometty’s Watson wants to read those transactions, then warn you that an upcoming tropical storm will kill your production schedule. (It just bought a big chunk of the Weather Company’s digital assets.)
Instead of software that tracks how many prescriptions were sold, Watson wants to tap into that data, then help individual patients figure out if the drugs they use are really working. (It’s working on that with CVS.)
In fact, Rometty thinks that transforming medicine with Watson is her “moonshot” project. She dreams of smart computers saving lives.
Microsoft is also working on machine learning, with slightly different takes. Microsoft is focused on smarter “productivity” apps to help you at work.
Oracle is starting to have its own take on machine learning. Ellison demonstrated something it calls a “just in time learning system” that helps sales people learn about competitive products before they pitch a product.
Ditto for Amazon which launched a machine learning service in April.
None of these approaches to the burgeoning cloud computing market are better or worse than the others, just different enough to apparently keep Oracle from competing head-to-head with IBM these days.
And the numbers tell at least part of the story. IBM said in October that its cloud services are on an annual run rate of $US4.5 billion. Oracle CEO Mark Hurd tweeted last week that Oracle’s cloud business is at annual run rate of $US2.4 billion.