Oracle Likes Cloud Computing After All

Larry Ellison angrily dismisses suggestions that Oracle’s business will be harmed by the rise of cloud computing. Many misinterpret Ellison’s remarks to mean he (and by extension Oracle) thinks cloud computing is a dumb idea that Oracle won’t pursue. We are now learning that Oracle does, in fact, intend to pursue cloud computing. But we're also learning that Oracle's strategy is more limited than those of IBM and Microsoft, its large-vendor competitors.

I attended the San Francisco Oracle Cloud Computing Forum on February 25, 2010 -- one of 51 such events Oracle is running worldwide. Oracle described a strategy that is pragmatic for Oracle, but not well-articulated and missing key pieces.


Oracle’s approach to cloud computing has three major parts:

  1. If you want an internal, or private, cloud, Oracle will sell you the hardware and/or middleware to build it.
  2.  If you want to use Oracle’s software on a third-party cloud, Oracle supports Amazon Web Services and Rackspace Cloud today, and will support other clouds in the future.
  3. If you want to rent rather than own Oracle’s business applications, Oracle will provide those apps under a hosted subscription model.

This strategy seeks to adapt Oracle’s current business modes and products to the concepts of cloud computing. To Oracle, internal clouds are a new data-center architecture, public clouds are a new sales channel, and cloud-based applications (SaaS) are new business terms. There’s no new operating system (a la Microsoft Windows Azure), no new development model (a la Salesforce’s, no commitment to provide public cloud platform services (a la IBM and Microsoft).

Oracle described a two-step process to move from today’s typical architectures to internal clouds that makes sense (see figure) -- and starts with adoption of Oracle's "grid" products. This progression makes sense, but not as easy as Oracle makes it sound. To be fair, most vendors understate the difficulty of transition from today's data centers to internal clouds.

In my opinion, the most innovative aspect of Oracle’s strategy is the Oracle Private PaaS (platform as a service). This offering is Oracle’s database and the middleware components most commonly used for development wrapped in a design concept for cooperative development. Using Oracle Private PaaS, application development and delivery pros create shared application services, components, application patterns, a self-service portal, and automated deployment routines. Business analysts then assemble, deploy, and manage their own apps--primarily departmental and temporary apps. This is a great idea, but it looked to me like Oracle's offering required too much "IT assembly" to be practical for many organizations.


Oracle’s two steps to internal clouds are clear enough (if simplistic), as is its understanding of the pros and cons of today’s cloud computing offerings. But the next level of detail at the session was messy. The problem: Oracle was selling only gear and software kit for cloud computing – not solutions. Customers have to work too hard to get to a solution. And there are too many choices without enough context. Why should customers start with Oracle RAC instead of Oracle VM and WebLogic Virtual Server Edition? That’s just one question, there are many other choices built into Oracle’s strategy.

Oracle spent very little time on its hosted applications at the session. These products are obviously solutions rather than stuff to build solutions. But Oracle has not created SaaS offerings from its huge portfolio of enterprise applications. Oracle CRM On Demand is its only multitenant SaaS product.

Creating more SaaS products will help simplify Oracle's offerings for customers. But Oracle also needs to define packages of its hardware and software components -- similar to IBM’s notion of a “cloud in a box” and/or Microsoft’s Business Productivity Online Suite (BPOS).


Lastly, Oracle's cloud computing strategy isn't complete yet.

The biggest gap in my opinion is Oracle’s cloud-unfriendly licensing. Oracle still measures software in units of hardware – cpus or chip cores. This means that when you deploy Oracle software to one virtual machine on an 8-cpu server, you must pay for 8 licenses. Oracle’s Amazon EC2 prices equate 4 EC2 VMs to 1 cpu for the Oracle Database Standard Edition. This isn’t VM-based licensing; it is a formula. Without VM-based licensing, variable costs based on resource usage will be impossible. And usage-based costs are a crucial reason to consider cloud computing.

And where are the assessment and migration services from either Oracle or its partners? As discussed, the real issues in migrating from current data centers to internal clouds are organizational, cultural, political, and knowledge gaps. Customers will need guidance and help in making their transition.

Also, Oracle discussed a new Assembly Builder to simplify cluster and cloud deployment at yesterday’s event, but many cloud-management tools were labeled "coming soon." The concept demos shown at the event suggested to me that Oracle must do a lot more integration of its development and operational-management tools to provide its cooperative development vision. For ISVs, there's no billing services in the strategy.And Oracle doesn't support VMWare's popular vSphere virtualization software.

Lastly, Oracle's strategy doesn't address public clouds strongly enough. I believe that the transition to internal clouds will be so difficult that most organizations won't make it. Rather, they will look to providers of various descriptions to gain benefits from cloud computing. Oracle's strategy doesn't really address this scenario. It has partnerships with Amazon and Rackspace, but treats these partners only as channels, as discussed above.

Why? I believe that cloud is a revolution for Oracle, IBM, SAP, and the other big vendors with direct sales forces (despite what they say). Cloud computing has the potential to undermine the account-management practices and pricing models these big companies are founded on. I think it will take years for each of the big vendors to adapt to cloud computing. Oracle is just beginning this journey; I think other vendors are further down the track.

If you want to get the story from Oracle itself, register for one of the 51 events it is holding around the world. You won’t yet find materials describing the strategy on, although the Oracle Cloud Computing Center page has useful information.


Application Development

The creative person who works as an application development should always think “out side the box”. It has always been seen that diverse organizations have diverse application demands. While creating the software we should always keep the origination and the business in mind.

re: Oracle Likes Cloud Computing After All

John, nice analysis. On your last point of Oracle not strongly addressing the the public cloud space, it seems to me as if they're purposefully trying to keep the focus on the private cloud to better leverage the $7B+ investment they just made in Sun. Do you see it the same way?

I'm also surprised at the lack of investment in SaaS apps. I remember seeing Anthony Lye announce Oracle CRM 15 with the Social apps in 2007. It looked like a lot of great functionality and they seemed to be talking it up as if there was a lot more to come. And last year it looked like they may be priming the market for more apps on demand when they revised the pricing model. But alas, not so much. Any insights as to why?

re: Oracle Likes Cloud Computing After All

Looks like my link to my Anthony Lye post didn't come through via html. So here it is:

re: Oracle Likes Cloud Computing After All

Glenn: I think Oracle is focused on private clouds because its corporate sales force can sell such things without a lot of disruption. Simple as that. Sun's servers and storage give Oracle's sales guys more stuff to sell. This was primarily what I meant by saying the cloud strategy is "pragmatic for Oracle." Please note that my colleague Stefan Ried disagrees with me -- he sees Oracle embracing public clouds but through partners. See his blog on this topic: I don't follow Oracle apps, and so don't know about the Anthony Lye business you brought up. Will look into it.

re: Oracle Likes Cloud Computing After All

John, I think Cloud Computing is a hype after all. Why? Because it is ridiculous to sell CC based on HW use. Even if the use of HW is very flexible and you get what you need, CC ought to mean to pay only for benefits that you get per user and in terms of service level. Our Papyrus Platform deployment is more like Cloud Computing because we charge one base fee for the system and then customer pay per user regardless of how many server nodes are deployed in the background to get to the SL needed. Only where user independent services are used we charge per server. We also don't charge for archive nodes, just for the server that indexes and stores content.

How would a business know what it really has to pay if it does what a CLOUD is all about - distribute its applications across a number of providers.

Last but not least: A Cloud would have to be about a new paradigm to create composite applications and I don't see any of that happening at Oracle or Amazon. It is all still about complex programming for integration.

Please focus more on realistic benefits for the user and not just on what vendors are selling.

re: Oracle Likes Cloud Computing After All

Thank you for your comments, Max. I'll take them into account when analyzing the question of "why cloud?" My purpose in this blog post was more narrow.