In the movie Ironman 2, Larry Ellison makes a cameo appearance as a billionaire, playboy software magnate. It is a role he knows well. He is playing himself -- chief executive of Oracle Corp, one of Silicon Valley's most enduring, successful and flamboyant figures.

At age 65, he is undertaking one of the biggest challenges of his career, and it's not playing Hamlet on Broadway. Oracle, the company Ellison founded three decades ago and built into dominant force in the software industry, is making a go at hardware with the acquisition of money-losing Sun Microsystems.

This is not entirely unlike MIT deciding to field a competitive football team, but Ellison being Ellison, he could not be less worried. We have a wealth of technology to package into systems, said Ellison, who won the America's Cup in February. I see no reason why we can't get this to where Sun under Oracle should be larger than Sun ever was.

In a rare interview he discussed his turnaround efforts at Sun so far, revealed plans to purchase additional hardware companies and detailed new products that will launch in the near future. And he did so with his usual in-your-face style -- heaping all manner of abuse, for example, on Sun's previous managers.

During the 1990s, Sun prospered by selling high-end computers at top dollar to large corporations and dot-com start-ups. Its business peaked in 2001, then slid with the collapse of the Internet boom and never recovered, though the company is still widely respected for its technological prowess and the brain power of its engineering staff.

Sun came into play in November 2008 after IBM CEO Sam Palmisano made an overture to buy it. Oracle, which had been strictly a software maker, unexpectedly jumped in to outbid IBM by just 10 cents a share, paying a total of $5.6 billion in cash.

Now Ellison says he is going to rebuild Sun's hardware business by using a strategy that helped IBM prosper in the 1960s: Selling computer systems built with standardized bundles of hardware and software.

Plenty of skeptics doubt Ellison can pull it off. Sun lost $2.2 billion in its last fiscal year as an independent company. Conventional wisdom holds that he will end up divesting the company's hardware business.

Ellison has a pretty good track record when it comes to predicting where the industry is headed. Besides innovating the wildly lucrative relational database that bears Oracle's name, Ellison was quicker than most in creating software that works with both Internet technology and the widely used Linux operating system.

He also started buying up smaller software makers in 2003 when critics said his consolidation strategy was doomed to fail. It hasn't. People have lost a lot of money second guessing Larry about IT strategy, said Dave Roux, co-founder of Silver Lake, the world's biggest private equity firm focused on technology, in which Ellison was an original investor.

He's a very thoughtful and reasoned observer of the big tectonic forces that kind of go rippling through the industry, said Roux, who worked for Oracle before setting up Silver Lake.

Ellison has maintained his status as the leader of a powerhouse in the topsy-turvy, protean technology world. IBM, which pioneered business computers, nearly collapsed in the 1990s, but then recovered as it aggressively expanded in services and software. Ellison's close friend Apple's Steve Jobs, was forced out of Apple, only to return a decade later to resurrect his company with the iPod. Meanwhile, Google has replaced Microsoft as the 'ubertech company' and occasional villain.

Although his products are used by businesses only and not nearly as recognizable as Apple's Macs or Google's search engine, they've made Ellison the world's sixth-richest man, worth an estimated $28 billion, according to Forbes. Oracle counts the bulk of the world's major corporations as customers, and the company's market value now tops that of Hewlett-Packard, the world's top maker of personal computers.

STARING AT SUN

Ellison says he has already stopped the carnage at Sun, less than four months after the sale closed in January. Their management made some very bad decisions that damaged their business and allowed us to buy them for a bargain price, he told Reuters. He added that he expects profit from Sun's operations to boost Oracle's earnings in the current quarter, which ends May 31.

The integration has proceeded swiftly, says Ellison, because a protracted antitrust review in Europe gave Oracle time to draw up an exhaustive plan for resuscitating Sun. In typical Ellison fashion, he took a hands-on approach to the integration, choosing to meet directly with technical managers at Sun as often as four days a week to diagnose its problems, rather than delegating the work to underlings.

Mark Barrenechea, a former Oracle executive who used to sit in on weekly engineering meetings with Ellison and is now CEO of specialty computer maker Silicon Graphics International Corp., says this is what Ellison does best. He doesn't write the code. He doesn't solder resisters onto motherboards. But he understands how all the pieces fit together and how he wants the building to look, said Barrenechea.

When Ellison started scrutinizing Sun, he says he found plenty wrong: lots of waste among the billions of dollars that had gone into R&D in recent years. The hardware maker had cut back the sales staff that sold its most profitable products, including its business computers and storage equipment, causing sales and earnings to decline. It operated an antiquated manufacturing and distribution system. It regularly sold hardware and software at a loss, sometimes losing more than $1 million on a single deal.

Oracle is much more disciplined on the financial side. Much more detail oriented, said John Fowler, head of Oracle's new hardware division and the most senior Sun manager to survive the acquisition.

Ellison says he learned that Sun's pony-tailed chief executive, Jonathan Schwartz, ignored problems as they escalated, made poor strategic decisions and spent too much time working on his blog, which Sun translated into 11 languages.

The underlying engineering teams are so good, but the direction they got was so astonishingly bad that even they couldn't succeed, said Ellison. Really great blogs do not take the place of great microprocessors. Great blogs do not replace great software. Lots and lots of blogs does not replace lots and lots of sales.

Schwartz declined comment as did Sun co-founder and former Chairman Scott McNealy.

At the start, Ellison shut down one of Schwartz's pet projects -- development of the Rock microprocessor for Sun's high-end SPARC server line, a semiconductor that had struggled in development for five years as engineers sought to overcome a string of technical problems. This processor had two incredible virtues: It was incredibly slow and it consumed vast amounts of energy. It was so hot that they had to put about 12 inches of cooling fans on top of it to cool the processor, said Ellison. It was just madness to continue that project.

More infuriating, says Ellison, is that Sun routinely sold equipment at a loss because it was more focused on boosting revenue than generating profits.

The sales staff was compensated based on deal size, not profit. So the commission on a $1 million sale that generated $500,000 in profit was the same as one that cost the company $100,000, he said. The sales force could care less if they sold things that lost money because the commission was the same in either case, he said. Ellison added that Sun also lost money when it resold high-end storage equipment from Hitachi Ltd, storage software from Symantec Corp and consulting services from other companies. Oracle is ending those deals.

Ellison has been pruning Sun's line of low-end servers, an area where it lost money as it tried to compete in volume with market leaders Hewlett-Packard Co and Dell Corp. Sun never gained enough share to make it price competitive, but Ellison says that Sun became so desperate to goose the unit's revenue the company would pay a fortune to charter planes during the last two days of a quarter so that it could book extra sales for the period.

REORG OF A SALES FORCE

In recent years, Sun outsourced much of its sales, counting on resellers to promote its products. That's anathema at Oracle, which employs 22,000 salespeople and 11,000 software consultants who work directly with its largest customers.

One casualty of that strategy was StorageTek, a maker of tape storage systems for mainframe computers, which Sun bought for $4 billion in 2005.

Astonishingly they laid off all the sales people and they laid off all the field service people. They just got rid of them all, he said. Guess what? Sales dropped. It's breathtaking!

The result: Sun went from being tied with IBM as the market leader in 2005 to No. 2 in 2009, when its $474 million in sales were about half those of IBM, according to data from IDC.

On the day the acquisition closed, Oracle advertised openings for some 2,000 hardware sales reps, which would almost double Sun's previous sales force of 3,000.

They have their work cut out. Sun's share of the server and storage markets slipped last year as the company's pending sale raised uncertainty about its future. It slid from second to third place last year in the $13 billion UNIX server market behind IBM and HP, while sales of its storage gear plunged 27 percent, according to IDC.

Now Ellison is looking to win back customers by boosting investment in Sun's key technologies -- including the SPARC microprocessor chip, the brain of its elite server line; the Solaris operating system that runs those computers, and the MySQL database as well as the widely used Java computer language.

It's picking the Sun technologies that are commercializable and focusing on those, and ignoring those that are not. They are just science projects, said Ellison.

LESSONS LEARNED

As he enters the hardware business, Ellison is looking to repeat the success he had with his last big strategy shift -- an acquisition spree that began in June 2003 with a hostile takeover bid for PeopleSoft, a maker of programs for managing human resources. Prior to that, Ellison was known to mock advisers who recommended that he consider acquisitions, saying we write software, we don't write checks. But he changed gears after more than a decade of unsuccessful attempts to design his own business management software, an area where Germany's SAP AG was rapidly consolidating its lead.

At the time there were few examples of successful acquisitions in the software industry, let alone hostile deals, but the purchase proved a winner, and Oracle has since spent more than $35 billion buying some five dozen software makers.

He said it's going to get harder and harder to build on our business because we have such a large market share. When everybody else was standing on the sidelines, he went deep and he went long, said Howard Anderson, a lecturer at the MIT Sloan School of Management.

Oracle quickly cuts duplicate costs and waste at a newly acquired company, boosting its margins and absorbing its technologies. Meanwhile it cross sells its vast line of business software to an acquired firm's existing customers while generating revenue from maintenance services for the programs they already use.

Shareholders have benefited. Oracle's stock price has nearly doubled since Ellison began his shopping spree, compared with a 21 percent gain in the S&P 500 Index.

DEAL TIME

Ellison said he will expand further in hardware, adding providers of semiconductor, storage and server technologies to his list of acquisition targets. We'll buy in all areas of our business.

He is looking for technologies he can incorporate into what Oracle is betting will be its next line of blockbuster products: gigantic appliances made up of servers, storage equipment, networking gear and software working together to handle specialized tasks.

Everything is tweaked at the factory for optimal performance in the data center. We don't want customers making those decisions. They never quite tune it the way we would, said Oracle President Charles Phillips.

This approach traces back to Ellison's strategy for building his software business. He has spent two decades telling customers that they should buy pre-integrated software business management suites from Oracle, rather than dealing with myriad vendors. His mantra: best of breed is good for dog shows, but not software.

Where we think we'll make our money -- where we think we're able to differentiate ourselves from IBM and everybody else -- is by building complete and integrated systems from silicon all the way up through the software, all prepackaged together, he said.

His first appliance was developed as engineers found that hardware bottlenecks slowed the pace at which computers could process information stored in Oracle's database software. So the company forged a deal with Hewlett-Packard to jointly develop an appliance known as the Exadata database machine, whose chief components were Oracle's database software and an HP x86 server.

Ellison introduced that first appliance in 2008 with the intention of developing other machines with HP. But when Oracle had a chance to buy Sun, Ellison ditched HP and entered the hardware business on his own. Last year he unveiled Exadata, version 2, based on Sun technology, which he says initially has nearly $1 billion in potential sales and will likely have annual sales measured in the billions of dollars within a few years.

The machine costs more than $1 million, stands over 6 feet tall, is two feet wide and weighs a full ton. It is capable of storing vast quantities of data, allowing businesses to analyze information at lightening fast speeds or instantly process commercial transactions.

But rivals remain unimpressed. They say that Exadata does not provide customers enough flexibility to customize the systems to meet their own needs.

It's got a very narrow band of applicability. It's a one trick pony that has no ability to integrate, said Dave Gelardi, vice president of systems strategy at IBM, which sells more than 20 types of computer systems optimized for specific tasks.

IBM also has the world's largest technology services organization, which generates billions of dollars a year in fees by helping companies build customized systems. We don't say 'The answer is Exadata. What's the question?' We operate from the perspective 'What is the business problem you are trying to solve?' says Gelardi.

Hewlett-Packard and Dell followed IBM's example of building a large services organization over the past two years with their acquisitions of EDS and Perot Systems. Ellison says he's not going down that path. While Oracle will maintain a relatively small force of consultants, his strategy is to do most of the integration work in the factory, based on the Exadata model.

THE NEXT STEP

He plans to unveil at least two new appliances in September, when the company holds its annual Oracle World customer conference in San Francisco, a gathering that had some 40,000 attendees last year, making it the world's biggest tech users conference.

The machine that will likely get the most attention is one that will run Oracle's yet-to-be-released next-generation suite of business management programs, dubbed Fusion Apps, which will compete with SAP's existing software line.

That appliance will include servers, networking equipment and storage gear along with a database, middleware and the Fusion Apps programs. Oracle has spent five years and billions of dollars developing the software, which helps businesses manage accounting, human resources, sales and marketing as well as more specialized tasks for companies in industries including financial services, telecommunications, healthcare and retailing.

Ellison says he will also show off an appliance that runs Oracle's so-called middleware -- the plumbing of a data center's software network, performing tasks such as helping various computer systems communicate with each other.

His key rival in this area, too, is IBM. And that fight won't be easy -- or short.

But Ellison says he expects to see it through and has no plans to leave Oracle any time soon. They may kick me out, but I have no plans to retire, he said.

(Reporting by Jim Finkle; Editing by Jim Impoco, Ted Kerr and Steve Orlofsky)