id="info"

Cloud

Oracle lays out its vision for Sun: Will it work?

Oracle CEO Larry Ellison provided industry watchers with quite a preview of the big plan for Sun. Here's what he says Oracle will -- and won't do -- with Sun.

This is a guest post from Larry Dignan of TechRepublic's sister site ZDNet. You can follow Larry on his ZDNet blog Between the Lines, or subscribe to the RSS feed.

Oracle expects the Sun Microsystems purchase to be approved by the European Union and close in January with a formal outline of the master plan and financial outlook to shortly follow. In the meantime, Oracle CEO Larry Ellison provided industry watchers with quite a preview of the big plan for Sun.

As far as the vision for Sun, it has been common knowledge that Ellison wants to provide integrated hardware and software. Think T.J. Watson's IBM. Or Apple for enterprise-class gear. However, a bevy of questions loom. How serious is Oracle about hardware? What will Oracle do with Sun's commodity server and storage businesses? Can Oracle play the market share game (Sun is getting crushed)? And should Oracle even attempt to play the market share game?

Ellison addressed some of those questions on Oracle's fiscal second quarter conference call, which followed a strong earnings report. While analysts on Friday were still mixed on the Sun purchase, they seem to be coming around to Ellison's point of view. Here's the plan for Sun and what Oracle will-and won't-do with it, according to Ellison.

Sun won't play for market share. Ellison was pretty blunt about Sun's prospects as a massive server player. Sun doesn't have the scale and frankly getting it isn't worth the effort. Ellison said:

One thing we recognized that Sun just really does not now and is never likely to have the volume to compete in the high volume, low margin business of just selling an Intel server with Windows on it or Linux on it one at a time. We think that high volume, low margin business is a good business as long as you have high volumes. That is something that Dell and HP are very good at and we are going to avoid that business.

Sun will go with high-margin, high-performance hardware where the secret sauce is software integration and raw speed. Ellison said "we are pursuing the high value, high performance market." For instance, large systems like the SPARC Solaris M9000 (right) will continue to be enhanced and naturally they will be tuned to run Oracle databases at breakneck speeds.

If all of this sounds downright mainframe-ish well that's because it is. Not only does Ellison want to emulate T.J. Watson's IBM it wants to punch the present-day IBM in the gut (the core mainframe business). Deutsche Bank analyst Tom Ernst said "Oracle's strategic focus to compete and displace IBM's mainframe business in the high-end of the market."

Oracle will play the cloud game with the aim of growing share in the data center. Ellison is adapting Oracle's cluster pitch to a cloud computing world. Ellison said:

We are also going to be building clusters of industry standard machines and SPARC machines. Those clusters are now called private clouds. That is the more fashionable term for clusters. We are using our software, our operating systems, both Solaris and Oracle Enterprise Linux, our virtualization, the ability to dynamically allocate and de-allocate resources which is essential for cloud computing as well as the integrated networking and integrated storage to deliberately complete private cloud to our customers. Our customers will be able to buy high end SMP machines, high performance and high value for high-end private clouds with all of the pieces including processing, storage and networking integrated together with Oracle/Sun software.

Exadata is just the beginning of integrated enterprise machines. Oracle executives said Exadata demand is off the charts and the company is having trouble manufacturing enough of them. The master plan for Sun is to use it to deliver middleware and application machines. If Ellison is correct, you won't buy ERP buy itself. You'll buy ERP on a machine tuned for it. Ditto for analytics. Ditto for middleware. Ditto for cloud operating systems.

Here's the big vision, which would stink for systems integrators.

Our overall strategy going forward as I mentioned earlier, is not to sell these individual industry-standard components on their own but rather group them together into machines like Exadata where we have processors, networking, storage, storage software, database software, our Oracle Enterprise Linux operating system, all a complete database machine for both transaction processing and data warehousing.

We think that makes it much easier for the customer. They don't have do all the system integration. They don't have to buy a bunch of parts and glue them together but instead they buy the box. It is a high margin product for us and a high value purchase for them because they don't have to spend a lot of money on system integration. We think that is the way customers are going to go forward as they build their data centers; not buying components but buying systems like Exadata.

One of the big reasons we bought Sun is that we wanted to apply that same strategy to middle ware, applications and to the operating systems themselves where we are not going to sell operating systems just for an individual computer but we are going to sell the next generation of Solaris is going to be a cloud addition to Solaris where it manages a group, a cloud, a cluster of these computers that we sell together as a unit. That is highly differentiated, high margin for us and no systems integration required for the customer. How big is that business? We think that is what the computer business is going to look like for large customers going forward.

If Oracle executes, Ellison reckons that Sun can differentiate itself from the likes of IBM, HP and Dell and make gobs of money too. Oracle reiterated its plans to deliver $1.5 billion in accretive operating income from the Sun purchase.

The larger question: Do you buy this vision? Analysts seemed to be coming around, but not completely. One big issue: Can Oracle get the hardware supply chain right? Indeed, Oracle had trouble manufacturing enough Exadata boxes to meet demand. JMP Securities Patrick Walravens said in a research note:

We like Oracle's vision of using Sun's assets to deliver unified solutions to customers to both power customer data centers and cloud provider data centers. With Sun, Oracle will have most, if not all, the tools needed to provide a full cluster, including processors, networking, storage, and database software. One area concern with Oracle's vision may be capacity constraints. Oracle will need to ensure it has both the hardware production capabilities and supply chain capabilities to make sure it is able to meet customer demand.

In other words, scale matters still and it's managing a hardware business is vastly different-and less lucrative-than consolidating the software sector and collecting maintenance revenue.

Other analysts such as Brad Reback at Oppenheimer question the long-term prospects of buying Sun, but generally seem more receptive of the purchase. Like his previous acquisitions, Ellison's chore will be proving skeptics wrong.

Here's a video of Ellison's big Sun plan:

Editor's Picks