Oracle: Our cloud to be cheap as open source

Oracle is making a push as a cloud provider, two years after its chief executive, Larry Ellison, dismissed the concept as mere marketing hype.

"Cloud computing is not only the future of computing, it is the present and the entire past of computing... All it is is a computer attached to a network," Ellison said in a speech in 2009. "Our industry is so bizarre. They change a term and think they've invented a technology."

But as more businesses have adopted the cloud, Oracle has warmed to the approach. It has developed private cloud hardware systems — Exalogic and Exadata — based on Sun technology, and Ellison launched Oracle Cloud Services at Oracle OpenWorld in October, saying, "We need a cloud."

Tyler Jewell, head of strategy for Oracle Cloud Services, talked to ZDNet UK about the company's nascent cloud and how it hopes to attract small businesses that, in the past, have been too "intimidated" by Oracle to use its products. In particular, he described how Oracle expects its services to end up as cheap to run as those based on open-source technology, such as OpenStack.

Q: What are Oracle's ambitions for its cloud service? Who do you want to see using it?
A: There's a couple of different categories of customers that are well suited to this. Pretty much the entire Oracle customer base, [especially] customers that have needs that require instant provisioning.

There have been a lot of projects where customers would like to use Oracle technology, but haven't been able to because they've had to go through IT and it would have been a very complicated process, and so they end up going with open source or OpenStack and the integration becomes much more of a problem because they have that urgent need. The cloud opens up a whole new class of applications.

On the second part, there's a whole group of [small to medium-sized businesses] that wouldn't think about engaging with Oracle because the model for licences, support and the pure breadth of our stack is probably a little bit of overkill or maybe a bit intimidating to them. The consumability approach that we're taking with this [means] we can open up the technology to a broad group of companies that wouldn't even consider it. I see this as a way into the small ones.

We are structuring our pricing in such a way that if you're a company and you're looking at a pure open-source stack because you think it's free and because you think it's cheaper, you have to look at the total cost of owning that, and it's non-zero.

We think we're going to be equal to or cheaper than maybe that [open-source] equivalent, but now you are on a tested and proven WebLogic database. If price is the only concern out there, there's going to be a lot of reasons for them to consider using our cloud instead of an open-source stack.

If price is the only concern out there, there's going to be a lot of reasons for them to consider using our cloud instead of an open-source stack.

There's [also] a whole class of third-party web service providers, like Salesforce.com, that provide extensions, and I think there's a whole as-a-service market coming along, like logging-as-a-service or buttons-as-a-service. These groups provide a lot of value to the lifecycle to the whole of the system.

Oracle's never really had a way of partnering with that. You've either had to build your software on the Oracle stack or not. I think [cloud] opens up hundreds of companies that are pure web companies that have a way of doing business with Oracle.

When it comes to infrastructure-as-a-service clouds, do you think Oracle could move to building a product such as that offered by Amazon Web Services?
No. I just don't see us going down that track. The problem with going down the stack is the only thing to compete on there is price, and it's just a race to the bottom. It's unclear where the value is. Oracle is focused on...

...really meeting the needs of IT application owners and business owners and that's where our expertise is and that's where we can add value.

In the past we've seen things such as Oracle linking into Amazon Web Services. Do you see it happening with the cloud, where I can structure an application so it's in the Oracle platform-as-a-service, but I can reach out and consume other clouds as well?
I sure hope so. I don't know if you need to reach out to other clouds to get scale. We have the elasticity built in.

The value of integrating with other clouds is that people are going to have a lot of different services, and they are going to choose to host them in different places. So, what you really have to do is make it simple for people to link their clouds together, so they have a common business process.

I don't know of many people who are doing true cloudbursting, [where] I have an environment in one place and I'm going to, all of a sudden, on-demand provision something else and link it together. There's a lot of synchronisation and there's a bandwidth pipe limitation there.

Think about what it takes to truly do cloudbursting. You have to have a total transactional data sync between any repository that you've got. If you have any amount of data getting into your database, that's going to be a hard thing to synchronise. The other thing is you have to support wide-area reprovisioning, so you can relocate all the sessions from one environment to another. I don't know of many IT guys who are thinking of that.

What I do see are a lot of them saying they want to have a blended cloud environment — a disaster recovery strategy where they have one cloud and another hot cloud standing by in case a disaster or something were to go wrong. Oftentimes, their premises are going to be the primary and the Oracle Cloud could be the hot standby, so all they have to do is get a reasonable amount of data over and then flip the switch.

I believe there's only going to be room for three big clouds in the world, and just on the basis alone of the ambition we've got, Oracle is going to be one of those three.

What can we expect to see in terms of innovation and forward development? How big is it going to get?
I'd be surprised if every Oracle customer doesn't try our cloud at some point. We have 305,000 customers around the world.

We're the only cloud that has a complete set of applications and a platform with it. We're the only vendor that does that, so we're going to be one of a kind. I believe there's only going to be room for three big clouds in the world, and just on the basis alone of the ambition we've got, Oracle is going to be one of those three. On that basis, we're thinking pretty big.

That implies a lot of capital spending on datacentres. How much will it cost?
It's not a cheap capital proposition, for sure. There's a lot of hardware, but you know what it is? It's still cheaper than our customers having to do it themselves.