Oracle Cloud: 5 Big Challenges
Oracle now wants its customers to migrate to the cloud, but it may be too little, too late. Here’s why.
It was clear at Oracle OpenWorld last week that Oracle wants its customers to start migrating to the cloud — not just to test-drive it, but to make a long-term, production-workload commitment. Of course the cloud they should migrate to is the Oracle Cloud.
“The transition to cloud is a generational change. This is a big deal, as big a deal as when PCs showed up in the days of minicomputers and mainframes,” Larry Ellison claimed. He sounded grave. Oracle is not just creating alternative cloud options to its on-premises products. It’s inviting customers to take advantage of its new services and join the next generation of computing — before it’s too late.
In case you missed it, there’s now the Oracle Cloud Infrastructure as a Service, Oracle Platform as a Service, Oracle Software as a Service, and at least a dozen other services with the appellation Oracle Cloud attached to them. That would include Oracle Analytics Cloud, Oracle ERP Cloud, Oracle Planning and Budgeting Cloud, Oracle Customer Experience Cloud, and Oracle Higher Education Cloud, to name a few.
[Want to learn more about how Oracle sees the completion? See Larry Ellion: Oracle Set Sights On Cloud Competitors.]
As all of this unfolded last week in San Francisco, I was struck by how the figures on stage now urging customers to migrate, were the same onesdissing the cloud not so long ago. I had no doubt that their words were sincere. CTO Ellison and CEO Mark Hurd want very badly for Oracle to be identified as a legitimate cloud vendor. They want to see their business move into the cloud, and they hope fervently that their customers pick up on their cloud initiatives.
Never mind that in October 2009 Mark Hurd, then CEO of HP, shook his head disbelievingly after a bad experience at a gathering of CEOs. “Here I am talking about the cloud … and I got a lot of boos, um, after that. It started with the whole term ‘cloud,'” he recounted. He concluded that we need some other word than “cloud.” That kind of cloud leadership, along with Ellison’s derisive references to “water vapor” and “nitwits on Sandhill Road” backing cloud startups are ancient history now. Even CEOs and CTOs have a right to change their mind, although it would be nice for them to at least genuflect once toward acknowledging past mistakes.
But if you’re looking for a reason why Oracle might be nervous, you could consult Wired’s Oct. 12 article by Cade Metz, which begins: “HP, Cisco, Dell, EMC, IBM, Oracle. Think of them as the walking dead.” The chief explanation given is they’re getting flanked by new cloud services and don’t respond quickly enough.
You definitely shouldn’t believe everything you read in Wired. It’s been wrong before. But still, it’s a scary situation for a $38.2 billion company with stagnant revenues.
Oracle executives aren’t looking to Wired so much as looking over their shoulder for a pursuing Headless Horseman. They probably noticed that Amazon Web Services is already huge and still growing quite consistently. It’s also making it extraordinarily easy for “proprietary database users” to migrate into its cloud, where it will charge one-tenth of what proprietary database vendors do, or so it says.
That doesn’t mean Oracle database users are about to bolt. But it swings further open a door through which many Oracle customers have already passed (at least once). If their database vendor is telling them even their production database applications will one day be in the cloud, then test-driving an Amazon-based system isn’t such a far-fetched idea. That isn’t what Oracle means when it urges them to migrate, but that’s the prospect many of its customers have in front of them.
For that reason, I find myself thinking about the five challenges that Oracle will have in building up its basic infrastructure-as-a-service cloud business. I think about IaaS because it’s the building block that leads to other cloud services. Oracle will have to succeed there if it’s got a long-term future in the cloud.
Challenge 1. IaaS Pricing Is Low-Pricing
Nothing is less familiar to Oracle than the need to constantly evolve prices downward and participate in a low price/low margin game. Oracle, throughout its existence, has successfully sold upfront software licenses, then set about improving the software to offer customers a new version and a new license. In between version sales, there are those annual maintenance fees of 22% with a profit margin of 90%-92%, Oracle tells investors. Overall margins for Oracle in the fourth quarter of fiscal year 2015 were 37%, according to generally accepted accounting principles.
Amazon Web Services reported what was for it a handsome margin of 25% in its fourth quarter. I’m sure the AWS margin wasn’t that large during the capital-intensive build-out of EC2 and its related services, but AWS is now capitalizing on the highly automated, self-provisioning services in which it specializes. If necessary, it’ll go back to lower margins to ensure that its cloud services remain the pre-eminent ones.
Amazon isn’t literally a bargain-basement, low-margin retailer, but its culture is much closer to a retailer’s than Oracle’s. To match Amazon and those cloud service providers determined to compete with Amazon — Google and Microsoft — Oracle will have to foster a price-cutting mentality throughout its business model and throughout its ranks. Good luck on that one.
Challenge 2. Pre-eminence of Open Source
The major cloud suppliers have been heavily dependent on open source code and both use it and offer it in many guises and forms. Microsoft is the exception, but it’s got its own bottomless well of Windows software that it can draw upon. CEO Satya Nadella and Azure CTO Mark Russinovich understand quite well how they have to intersect with open source programmers and projects or else miss out on a large chunk of cloud business. Next-generation applications being built for use in the cloud start out with open source. They are sometimes completely open source component-based and scale out over many servers without incurring software licensing charges, billing only for the cloud service time.