Softwar

by Matthew Symonds

  Ellison thinks that the sheer distraction of destroying Netscape and dealing with legal consequences damaged Microsoft more than is generally realized. He’s fond of saying that Microsoft “robbed the wrong bank.” He says, “Microsoft wasted a tremendous amount of moral capital, credibility, and time destroying Netscape. At Oracle we used to call Netscape the ‘heat shield.’ Microsoft was so busy burning Netscape to a crisp that they fell further and further behind in database technology. We have a big lead in server software, and I don’t think that they can catch us. Unlike Netscape, we’re protected by a huge, almost insurmountable technology barrier.”

  I couldn’t help feeling that there was at least a whiff of the hypocrisy that Ellison detests in the tear-jerking way that he talked about the crushing of poor little Netscape, while he seemed quite happy to try to snuff the life out of smaller competitors such as Ariba and Commerce One, which he derided as “features rather than software companies.” It’s a charge that nettles Ellison: “We’ve never seen anyone compete like Microsoft. Nobody. Microsoft very effectively and illegally threatened and bullied PC companies that dared to distribute Netscape’s browser. All the PC companies stopped distributing Netscape because they were afraid of Microsoft’s reprisals. Microsoft will go on breaking the law until someone stops them. They’re going after Real, using the same exact bundling techniques they used against Netscape. They’re bundling instant messaging to go after AOL. That’s how they compete: bundle, bundle, bundle. Why not? From their point of view, they’re getting away with it. I’ve never seen anything like this degree of abuse of monopoly power. IBM used to have a monopoly. But they never engaged in the kind of behavior that’s business as usual at Microsoft.”

  Okay, but if Oracle bundles its application server with its database, which is pretty much the case, and goes to war with BEA, which a lot of people see as an innovative company that has created a new market, what’s the difference between what Oracle’s doing and Microsoft? “First, we don’t bundle our application server with our database; we charge separately for our application server. Second, even if we did bundle our application server with the database we wouldn’t be breaking the law because we don’t have a monopoly in database. Under the provisions of the Sherman Anti-Trust Act, standards of behavior are different if you have a monopoly versus not having a monopoly. The law says you can’t use one monopoly to obtain another monopoly. Microsoft has a monopoly in desktop operating systems. Microsoft bundled its browser in their desktop operating system, and they succeeded in getting a monopoly in browsers. That’s express violation of our antitrust laws. Now Microsoft is bundling a media player in its desktop operating systems to get a monopoly in those. Microsoft is bundling instant messaging. See any pattern here? We have no monopolies; we don’t bundle; we don’t break the law.”

  I wondered whether Oracle, if it ever did get monopoly power, would behave more like Intel than Microsoft. When Andy Grove first realized Intel was on its way to having a dominant market share in PC processors, he made certain that everybody at Intel who dealt with the world outside understood exactly what the limits on their behavior should be. It didn’t stop Intel acting, at times, in a fairly brutal way toward competitors, but it had always accepted that the law was there and that it applied to Intel. Ellison laughs. “I think that obeying the law’s a really good idea. IBM and Intel managed their monopolies and competed aggressively without ever operating outside the law. But Microsoft is different. They understood the law. They had lawyers advising them. They’re smart guys. They’re not naive. But Bill’s an ideologue. I think he believes that the antitrust laws are wrong, so he doesn’t need to obey them.”12

  I had often thought that one of the reasons for Gates’s apparent willingness to flout the antitrust laws time and again was simply the arrogance and sense of inviolability that comes from unimaginable wealth. If anyone had a take on that, it should be Ellison. His answer was complicated. “Bill’s hyperrational side is in a losing battle with his emotionally fragile side. Bill goes out and methodically searches for good ideas to steal. That’s perfectly rational behavior. That’s made him very successful. But then, one by one, Bill starts to claim credit for the stolen ideas. He actually starts believing that they really were his ideas in the first place. He’s pretending to be someone else. He wants to be someone else. He can’t bear to see himself as Rockefeller; he sees himself as Edison. And he can get very childish and upset when other people don’t see him as the great inventor.” He tells a story of a dinner party at Andy Grove’s house during which Grove was asked whom he admired most in the PC industry. Grove replied that it had to be Steve Jobs: he had either invented or popularized everything to do with the PC industry. News of this eventually reached Gates, who called Grove and asked him to dinner. When the meal was nearly over, Gates finally let it all out: His feelings have been terribly hurt. Doesn’t Andy admire Bill?

  Ellison says: “Here’s this guy who built the most successful company on earth, but he’s still very unhappy with himself. He’s unhappy with how he’s perceived, not just by the public but by the people who are close to him: his primary partner, the guy who ran Intel. His behavior during the antitrust trial wasn’t so much recklessness coming from extreme wealth as self-destructiveness because he felt persecuted and unappreciated. Bill just wants to be loved. Who doesn’t? But historically America has never offered great love to people who accumulate vast fortunes. Just the opposite. If you’re rich, the best you can hope for is to make the transition from robber baron to philanthropist with a great deal of humility and a lot of check writing. Fair enough. That’s a perfectly reasonable and time-honored bargain. What else can you do with all that money, anyway? You can’t spend it, so you have to give it away.”

  America hasn’t yet “offered great love” to Bill Gates, at least inasmuch as he is now seen by most people as a latter-day Rockefeller, a ruthless bully, rather than the benignly grinning and blinking supergeek who built America’s most admired company. The government had been successful beyond all expectation in making its case against Microsoft, thanks to the courage of many witnesses, the courtroom skills of David Boies, and Microsoft’s own almost cavalier self-incrimination through both the testimony of its executives and the damning trail of e-mails.

  One consequence of the trial, I reckoned, was that the hardware makers had become braver about not being used as Microsoft’s pawns, and there were greater opportunities for Oracle to forge strategic partnerships with the likes of Compaq (before it was absorbed by HP) and Intel, as it was now doing. Ellison agreed up to a point. “Do you remember when Microsoft invited all those people to New York to get up onstage with Bill and say how much they liked him and how much they like working with Microsoft? It reminded me of Al Capone doing a similar thing in Chicago back in the thirties. It’s the ‘Be nice to the alligator and maybe he’ll eat you last’ theory of survival. Well, I’ve got a better idea. Let’s kill the fucking alligator before he kills us! If we fight back, maybe we’ll survive. Maybe we won’t kill the alligator, it’s a huge fucking alligator, but trying to kill it will improve our chances for survival. Being nice to this alligator will never work. If you think Microsoft’s competitors dislike them, you should talk to their ‘partners.’ Microsoft has bullied Intel and Compaq so many times they’ve lost count. Everyone in the industry is looking for an alternative supplier of software for Intel’s mass-market computers. Everyone’s thrilled about Linux. And I believe they’re looking to Oracle to provide some balance against Microsoft. IBM’s really not a big player in the software business except on mainframes, so it’s up to us. We’re the last, best hope for mankind in the battle against Microsoft and its plans for world domination.”

  To that end, Ellison was devising a strategy for Oracle to go into the operating system business. He believed that despite the hype over Linux and IBM’s enthusiasm for the open-source operating system, it would still be years before it was sufficiently hardened to do service in the corporate data ce
nter. For that reason, he was in discussions with Compaq’s boss, Michael Capellas, an ex–Oracle hand, to rework, and eventually take over, the immensely fast and robust Tru64 operating system, which had been developed over more than two decades by the best engineers at Digital, so that it would run on Intel’s new 64-bit processors.

  Ellison eventually wanted to see all the Intel-based computer makers selling certified servers preinstalled with Tru64 and Oracle’s database software. It would mean that Oracle would have preintegrated every part of the software “stack” from the applications at the top of the stack through the applications server to the database and, finally, the operating system at the bottom. “We’re going to build our next generation of database clusters out of a bunch of low-cost Intel computers. Our Intel clusters will be highly secure, totally fault-tolerant, and much faster and cheaper than anything else on the block. As for the operating system—don’t ask, don’t tell. You don’t ask what operating system a Cisco router is running. Why should anyone care what operating system our database cluster is running?”13

  It seemed that regardless of the final outcome of the Microsoft case, Oracle had already benefited from it. Even if Microsoft emerged constrained by only mild conduct remedies, Ellison seemed to think that Oracle could withstand any vengeance that Gates might want to exact. Rather than Oracle being on the firing line, Ellison believed that the companies that had most to fear were in the consumer and entertainment businesses. In particular, he thought that AOL Time Warner and Sony were about to feel the full effects of competing with Microsoft.

  Ellison frowned. “I think AOL is in deep trouble, but what do I know? I really don’t understand AOL’s business at all. To me it seems like a simple online service for people who don’t like computers. Whenever I prowl around their system, all I can find are a bunch of twelve- to fourteen-year-olds in chat rooms. Chat rooms are closely related to instant messaging. Both have been critical to AOL holding on to their customers. AOL chat and instant messaging is a closed network; they only allow you to communicate with other AOL members. That strategy works fine for a while, but it will fall apart over time. Imagine an e-mail system that only lets you e-mail other AOL members. No one would use it. As instant messaging matures into a standard service, AOL must open their system up or be isolated from all the other instant messaging users in the world. AOL has no good choices here. Either way they lose. Microsoft is bundling instant messaging with the next version of Windows. Instant messaging will soon become a standard service in business. There will be more instant messaging users outside AOL than inside AOL. It’s not a pretty picture. To make matters worse, Microsoft can underprice AOL whenever they feel the government heat is off. What if Microsoft priced MSN at half the cost of AOL for the next ten years or so? Then there’s the small issue of the upcoming competition from the broadband carriers—the phone and cable companies. Maybe they’ll squeeze out both AOL and MSN. Who knows? All you really need is a fast line from a carrier and Yahoo! Why should the broadband carriers share any money with Microsoft or AOL? AOL has a very uncertain future. Steve Case pulled off the deal of the century by buying Time Warner. AOL shareholders should give him a medal. On the other hand, Time Warner shareholders have a right to be very, very upset. And I think they’ll get more upset over time. What was Gerry Levin thinking?

  “X-Box [Microsoft’s game console] is likely to get killed by Sony’s Playstation 2 for the next couple of Christmases. However, Microsoft has great patience and endurance. They’ll be around year after year. Microsoft’s monopolies generate so much cash that they can blow billions in the game business and keep coming at you. That’s what Sony faces. Perhaps Microsoft won’t kill Sony’s game machine business, but they’ll be Sony’s number one competitor. And as games go online with broadband, Microsoft is likely to start building games that work with both PCs and the X-Box. The online game business is brand new, and that presents opportunities to a newcomer like Microsoft. Sony’s a great company, but Microsoft has unlimited financial resources. It’s no fun competing with a company that has an infinite amount of monopoly money—real monopoly, real money. So I think AOL and Sony are both vulnerable to competition from Microsoft. The barriers to entry into their businesses are primarily financial, not technical. Fortunately, the opposite is true for us. The reason we’re able to compete with Microsoft is because our technology is so hard to replicate. It took us more than twenty years to build our database. And our guys are every bit as smart as theirs. Sometimes I think that we’ll be their last competitor. The last one standing.”

  Hating and baiting Microsoft, software’s evil empire, is one thing, but I wondered whether Ellison could really get himself so worked up about IBM, a company he had admired all his life and was even now ready to call a “national treasure.” “Well there are two different IBMs—the IBM of Christmas past and the IBM of Christmas present. The old IBM, the real IBM, was created by Thomas Watson, Jr. That IBM was the greatest company in the history of the earth. When I grew up in this industry, IBM wasn’t someone against whom you competed; IBM was the environment in which you competed. But that was then and this is now. IBM never really recovered from the loss of Tom Watson. He was followed by a string of anonymous sales executives who just kept doing what IBM had always done: develop and sell mainframe hardware and software. As the center of gravity of the computer industry moved away from mainframes, IBM failed to adapt. The world changed, the industry changed, but IBM didn’t. Eventually IBM was forced to hire an outsider, Lou Gerstner, to come in and change the company. He abandoned IBM’s technology-centric past for a service-centric future.

  “IBM hasn’t developed any interesting software since the mainframe version of DB2 nearly twenty years ago. That’s a long dry spell. All of their great software is their old mainframe stuff: OS/390, VM, the mainframe version of DB2—all first-rate pieces of technology. The new IBM no longer develops much software; instead they go out and buy it. They bought Lotus, Tivoli, Informix, Rational, and so on. IBM’s products on UNIX and Windows are not close to being competitive. Their database on UNIX and Windows used to be called UDB [Universal Database] but they changed the name to DB2 to try to fool people into thinking it was the same database product that they had on mainframes. Now, DB2 on mainframes is a great product. DB2 on UNIX and Windows is a totally different product with a totally different architecture, and it’s a very distant third in the modern database business behind Oracle and Microsoft’s SQL Server.

  “Lou Gerstner turned IBM into a giant service organization. You have money; we have people. Want to trade? There is very little leverage in that. It’s really just a bigger version of EDS [a big systems integrator and hosting firm founded by Ross Perot]. But I don’t think building lots of oil and lube places to service your car is a very good idea. What if someone builds a car that only requires the oil to be changed every 100,000 miles? IBM’s service business model depends on computer technology remaining very complex and labor-intensive. IBM ran this series of ads showing a roomful of business executives saying: ‘Good grief, if this merger goes through we’ll have seven different platforms, fifteen different databases, fourteen different operating systems, a wireless project, and so on.’ Then you hear this woman’s voice speaking to the roomful of desperate execs, and the voice says, ‘And that’s when you know it’s time for IBM.’ In other words, computing is so enormously complicated that a normal business can’t hope to understand or cope with it, so you should call in IBM. I have a different idea. Use better technology that’s easier to understand.

  “Our strategy is to simplify our software so you need less labor to install and run it. You don’t need a big systems integrator if you don’t do much systems integration. IBM recommends that you buy lots of different applications from lots of different vendors. In fact IBM resells applications from SAP, Siebel, i2, Ariba, pretty much everyone I can think of except Oracle. Then IBM makes a bundle by selling you guys with glue guns to stick it all together. The old IBM’s strategy was c
reating value by creating intellectual property. The new IBM’s strategy is based on servicing complex systems; the more complex the system, the better it is for IBM. But if I’m right and the industry turns away from these labor-intensive, highly customized best-of-breed software systems, then IBM’s service business will slowly melt away, like a glacier.”

  The essence of Ellison’s hostility toward IBM is a form of intellectual snobbery. A company that was made great by its technology and applied engineering skills had lost its way and sold out. It was now run by sales and marketing guys who were cynically selling out the greatest name in American business. For the pragmatic Gerstner, remaking IBM as “the solutions provider with the deepest bench,” as his successor, Sam Palmisano, once described it to me, has been a triumphant success. But it’s easy to understand why Ellison sees the shift not only as an exploitable point of difference with Oracle’s main competitor in the database market and the aggressive reseller of its rivals’ applications, but also as a kind of betrayal. What most riles him, however, is the cavalier approach to facts that Big Blue routinely adopts in its marketing: “IBM went on the attack about six months ago. They sent out a bunch of letters to CIOs claiming that they were number one in database market share and number one in database performance. Well, they are number one in mainframe database market share, but on UNIX, Linux, and Windows we have five times the market share they have. Tons of surveys—Morgan Stanley, Goldman Sachs—support that. As for being number one in performance, well, I’m stunned that IBM would make such a completely indefensible and easily disprovable claim. The old IBM would never, ever have done that.”

  Although Oracle’s big philosophical battle with IBM is about complexity—the integrated Oracle suite versus the best-of-breed applications recommended by IBM and expensively glued together by Big Blue’s army of integrators—Ellison has decided that the first stage of the campaign against IBM will be “to go after their software technology products because their software technology isn’t what it used to be.” And it’s not just IBM’s database that’s in the firing line, but also its application server: “IBM’s got this thing called WebSphere. WebSphere is not a product; it’s many products: around a hundred separate products, I believe. It’s a bit confusing. IBM has gotten into this odd habit of using the same name for totally different things. DB2 is one name, but IBM applies it to two totally different products. The confusion is intentional. The WebSphere name takes the confusion to a new level. WebSphere includes a hundred different products. It’s the same naming scheme that George Foreman uses to name his kids: they’re all named George. Anyway, one of the products that falls under the umbrella name of WebSphere is IBM’s Java system. In a fit of creative license IBM put a statement on their Web site claiming that IBM’s applications server—WebSphere—runs Java twice as fast as BEA’s applications server. They didn’t publish their benchmark results, but they did put a copy of the benchmark program on their Web site. Now, we were ready to take IBM at their word because, after all, they’re IBM. They wouldn’t just make this stuff up. We had already tested BEA’s applications server and found that our Java was twice as fast as BEA’s Java, so we figured that we must be around the same speed as IBM’s Java. That was a little disappointing, but those seemed to be the facts.