Story from the Economist
“I AM very surprised. I have to think about it.” That was the initial reaction of Steve Ballmer, the boss of Microsoft, the world’s largest software firm. It was also the response of many in the computer industry to the news on April 20th that Oracle, another software giant, was paying $7.4 billion to buy Sun Microsystems, an embattled computer-maker and Oracle’s neighbour in Silicon Valley.
It was no secret that Sun, which never really recovered from the dotcom crash, had been searching for a buyer. But most industry observers had expected it to restart talks with IBM, another industry heavyweight, which had offered to buy Sun in March. Others thought Cisco, the biggest maker of networking gear, would make a bid, since it has recently started to move into computing hardware, too. Some even predicted that Hewlett-Packard (HP), the number one computer-maker, would finally get involved, after rebuffing Sun.
Why would Oracle, which makes most of its money from databases and business software, buy a hardware firm? A big part of the reason is the outsized appetite of Larry Ellison, Oracle’s flamboyant chief executive. His firm has spent around $30 billion buying 50 firms since 2005, among them software heavyweights such as PeopleSoft, Siebel and BEA. Buying Sun should help Mr Ellison achieve his goal of getting Oracle’s revenues above $30 billion by the end of this year.
There is more to the deal than that, though. For although Sun makes most of its money selling server computers and storage devices, it is not just a hardware-maker. It has always been a “systems” company, meaning that it sells tightly integrated bundles of hardware and software. And in recent years it has been beefing up its software business as companies have increasingly eschewed its pricey proprietary servers and bought cheap standardised machines instead.
Mr Ellison is keen on two bits of Sun’s software portfolio in particular. One is Java, a programming language that is the underlying technology both for many business applications and for software that runs on mobile phones. Sun never managed to make much money from it, in part because it wanted Java to be an open standard. But Mr Ellison may have different ideas. To him, it is “the single most important software asset we have ever acquired.” Sun’s other crown jewel is Solaris, its highly reliable operating system, which is often used as the platform for Oracle’s databases. More Oracle databases run on Solaris than on any other operating system, Mr Ellison notes. With control over both pieces of software, Oracle will be able to make them work together better.
This ability to integrate hitherto disparate pieces of technology, and thus make life easier for companies, provides further justification for the merger. For some time, Mr Ellison’s vision for Oracle has been to become the Apple of the enterprise, hiding complexity from customers, just as Apple does with its powerful but easy-to-use consumer products. Taking over Sun, he said this week, provides Oracle with all the pieces to put together systems that reach from “application to disk”. Oracle’s engineers are already brainstorming about how to build “industries in a box”—complete computer systems that come fine-tuned for, say, banking or retailing.
Yet there are other aspects of the takeover which Mr Ellison prefers not to mention. Top of the list is open-source software. In recent years Sun has put together a sizeable portfolio of these free programs written by communities of developers, hoping that firms using this software would then buy its expensive hardware. By buying Sun, Oracle becomes the world’s largest open-source company, prompting much debate among developers and users. There is particular concern about the fate of MySQL, a firm Sun bought for $1 billion in January 2008. It sells database software which is also available in a free, widely used, open-source version.
Next on the list is job losses. If Oracle’s previous acquisitions are any guide, Sun’s workforce, which has seen many waves of lay-offs in recent years, will be cut further, so that Oracle can meet its targets for the deal. Safra Catz, Oracle’s president, expects Sun to contribute more than $1.5 billion to Oracle’s bottom line in the first year and more than $2 billion in the second. Money, even more than the scope for tighter integration, also explains why Oracle is unlikely to ditch Sun’s hardware business soon, if at all. It provides a healthy stream of maintenance fees, about 40% of Sun’s sales.
Finally, the takeover is also a defensive move. Oracle did not want to let IBM get its hands on Java and Solaris, and felt it had to react to what looks more and more like a thorough restructuring of the computer industry. Since the early 1990s the industry has resembled a cake made of horizontal layers of technology, with each layer dominated by a few companies. Cisco, for instance, provided most of the networking gear. Sun and HP sold servers. Oracle was the leader in databases. IBM’s mainstay was services. SAP, a German giant, ruled in business software.
This structure is now collapsing as the industry’s heavyweights move into each other’s layers. HP bulked up its services division by buying EDS, for example, and has also moved more into networking. Cisco will soon start selling servers, and has formed an alliance with several smaller hardware and software firms to build, in effect, a data centre in a box. The industry is, in other words, going back to its past, when it was dominated by a few integrated companies that tried to do it all.
This is, in part, a consequence of the industry’s maturity: to keep growing, firms have to invade each other’s markets. In addition, customers increasingly prefer to buy integrated systems from one vendor, rather than doing the plumbing themselves. New technologies such as virtualisation and “cloud computing” are also blurring the boundaries between the industry’s layers. A server can now easily switch between being a computer, a storage device or a router (a box that directs network traffic). For computing to become a utility, which is the promise of the cloud, a data centre cannot be a hotch-potch of boxes cobbled together from different vendors, but must be tightly integrated.
Some analysts think the deal could accelerate this trend by triggering more acquisitions. IBM, for instance, may now be tempted to buy SAP, to pull even with Oracle and its big software portfolio. Cisco, which is sitting on $29.5 billion in cash, might take over some of its allies, such as EMC, a storage firm. Before Oracle’s competitors do anything rash, however, they should wait to see how well Mr Ellison’s firm can integrate Sun. So far Oracle has shown that it can make money by buying software firms, getting rid of overlapping products and unnecessary overheads, but otherwise leaving them much as they were. With the takeover of Sun, however, Oracle faces a more difficult integration. Will it enable Mr Ellison to help himself to a big slice of the industry’s cake?
