OOW17 - Oracle's cloud epiphany has come late, but it's no less genuine for that

Phil Wainewright Profile picture for user pwainewright October 4, 2017
Oracle's cloud epiphany at OOW17 has come late in the day, but don't dismiss it as bluster - this is a genuine conversion to the cause

Larry Ellison on stage at oow17
The prevailing view in Silicon Valley is that Oracle — a standard-bearer for the conventional mainstream in enterprise IT — is way behind the curve when it comes to the cloud. So the verdict on CTO and founder Larry Ellison railing against Amazon Web Services (AWS) at this week's OpenWorld conference in San Francisco has been that it's just bluster — Larry being Larry.

It doesn't help Oracle's case that Ellison's company has a track record of making outrageous claims for products it hasn't yet finished building. In fact, the company is literally a case study in Geoffrey Moore's iconic Crossing the Chasm series of business books on how to steal a march on competitors with this precise tactic.

This doesn't go down well in the cloud market, where it's the norm — with some exceptions — for providers to announce new products or features at the same time as they become available, rather than months in advance. But Oracle 18c, the new autonomous database that Ellison claims is dramatically faster and cheaper in the Oracle Cloud than equivalent workloads running on AWS, doesn't become available until December, and then only for analytics. Customers that want to run their mission-critical OLTP workloads on 18c will have to wait until halfway through next year. And there'll be waitlists for both varieties.

Not playing by the rules

So when Ellison cites benchmarks showing how much of a better deal 18c represents, he's not playing the game the way that cloud players understand the rules. He's using a classic ploy from the heyday of client-server computing, where by unveiling an upcoming release, incumbents hope customers will postpone their buying decisions until it becomes available. It's especially frustrating when the benchmarks may well have changed by the time customers can get their hands on 18c, but the ploy works in a market where buying decisions typically take months rather than days to close.

Adding to the sense of grievance is that while self-healing, autonomous cloud databases may still be a new thing at Oracle, the likes of Amazon and Google have been running this kind of database at scale for more than a decade. They long ago rejected Oracle as an old-school platform that didn't meet their needs, and they're not about to start taking lessons from an old-timer like Ellison.

On the other hand, while they may quite rightly feel they're experts in building and operating cloud-native databases, their own track record at building trust with the B2B enterprise market that Oracle still dominates leaves a lot to be desired. Some of Silicon Valley's own stars — Salesforce for example — still run on Oracle. And while it's true that cloud is growing fast, it's still a minority share of total global enterprise IT spend, a market that Oracle knows intimately. Ellison may be an old-timer, but his company's no has-been.

Speaking the language

The cloud-native generation would probably argue that it has been getting better at speaking the kind of language enterprise buyers understand. Conversely, Oracle is now becoming adept at speaking the language of cloud providers.

One of the most striking features of OpenWorld this week has been to hear Oracle spokespeople talking about what it really means to be a cloud provider. They're saying exactly the same things I've heard from cloud-native and SaaS providers over the past two decades of covering this space, but I've never heard it spoken with such authenticity at Oracle. Indeed I've often got into arguments with prior generations of Oracle product managers, putting exactly the same points to them that their successors are now expressing as if no one in the world ever realized these things before.

Here, for example, is Juan Loaiza, Senior Vice President of Systems Technology, explaining why customers expect a higher standard when they're relying on a third-party provider than when running their own infrastructure:

Part of the reason why it's difficult is, when you make something autonomous it actually has to be safer than the manual version. No one's going to buy an autonomous database that's less reliable or less secure than it is today.

Yes indeed, welcome to the cloud.

Building towards the goal

The stakes are particularly high in the case of transaction systems, because enterprises run mission-critical applications on them. But Oracle has been building towards this goal for a long time, says Loazia.

This is not stuff that we invented yesterday. This is stuff that has been around and has matured. We've worked with the biggest companies in the world to make this technology.

We've been doing this for decades, and we've built up this substrate of automation, high availability and intelligence in the database to the point where now we're ready to put the cap on it basically and say, 'OK now, we're going to provide you a fully autonomous database.' But it is a lot of work.

Of course there's bluster — an Ellison keynote wouldn't be complete without a good dose of competitor bashing. But don't be misled that there's nothing going on here. My take is that Oracle has finally engineered its platform to run in the cloud, and that's a huge step.

A grey colored placeholder image