Thursday, March 19, 2009 11:14 AM
commentary Stranger things have happened, but there are several reasons why IBM buying Sun Microsystems could, to borrow a phrase from former Sun Chief Executive Scott McNealy, be like two garbage trucks colliding in slow motion.
The Wall Street Journal reported that IBM is in talks to buy Sun for at least US$6.5 billion in cash, which would amount to about US$4 billion once Sun’s cash and marketable securities are taken into account.
On paper, the deal could make some sense: adding Sun’s server market share would give IBM more clout in its competition with Hewlett-Packard, IBM would get some software and intellectual property assets, and that price would be a nice premium for Sun shareholders disappointed with the company’s sliding stock price.
But given how directly Sun and IBM product lines overlap, there are no shortage of serious difficulties, too.
Let’s start with hardware. IBM already has four major server lines running a variety of operating systems–AIX, zOS, IBM i, Linux, and Windows–on three major processor families. IBM needs Sun’s Sparc processor and Solaris operating systems like it needs a hole in its head. It took years for IBM to break down some walls among its various server fiefdoms, but even now it has to reckon with complicated, often overlapping product lines.
Granted, Sun has, in its current Niagara and future Rock processors, some great intellectual property and expertise in designing multithreaded processors designs that can juggle a lot of tasks at the same time. But IBM would either have to adapt that technology to its own Power processors, a process that would take years, or embrace Sparc chips in its own line.
IBM or Sun could sell or license the Sparc line to Fujitsu, which has a line of its own and a Sun partnership. But the fact that Sun hasn’t done so on its own doesn’t bode well for Fujitsu’s enthusiasm for the idea, and buying Sun just to sell off one of its mainstay businesses erodes the market-share-grab rationale for the overall acquisition.
Sun also has a respectable line of x86 servers using chips from Intel and AMD. They aren’t a quantum level above the competition, though, and IBM already has a lot of in-house expertise with heavy-hitting x86 servers.
In storage, Sun made a big bet by buying StorageTek for tape drives that compete directly with IBM products. IBM might be able to consolidate customers in that market, but it’s not a big growth area. Potentially more interesting, though, is Sun’s Thumper line of x86-based storage devices, which have shown some life.
Software is another tough sell for the bean counters. IBM’s embrace of Sun’s Java helped cement its success on servers, but for Sun, Java is more about intellectual property, industry influence, and bragging rights than big money. IBM might well have “Java envy”, as McNealy quipped in 2004, but it can console itself with having a much larger Java software business in WebSphere.
The open-source connection in general is stronger. Both Sun and IBM have a history of both proprietary and open source software. IBM got an early edge through its embrace of the Linux operating system, support for the Apache server software, and founding of the Ecplise programming tool project, but Sun arguably has leapfrogged IBM with a more-open-than-thou philosophy under Jonathan Schwartz; Sun’s open source move now embraces its two biggest software assets, Solaris and Java.
More compelling for IBM perhaps is MySQL, an open-source database product widely used to power up-and-coming Web sites. IBM knows how to sell a database, but MySQL fits in a market where IBM’s DB2 doesn’t.
However, making money from open source software is a challenge, even if it’s a great way to appeal to developers and to needle Microsoft. So the appeal of Sun’s software business is much less direct than something that would contribute to IBM’s top and bottom lines.
Cloud computing, which combines hardware and software, is an area where Sun has some experience and some bruises; it announced a second attempt Wednesday to tackle the market for a general-purpose computing foundation that customers can pay for as needed. IBM has some experience in the area–including a history that extends to a cloud computing progenitor of decades past called time sharing–and overall, it’s hard to imagine that IBM is unable to do this on its own.
Sun and IBM have different cultures that could prove difficult to integrate. Sun, based in Silicon Valley, is an engineering-centric, free-wheeling company willing to try many ideas and see which ones stick. IBM is more conservative and driven by business concerns. Its bold moves often take years to pan out. Both companies share a passion for research and development, but how they bring that to market differs greatly.
Sun employees looking at the company’s troubles might well be happy to don blue Oxford shirts, at least figuratively. But it’s not easy to reconcile different procedures for allocating resources, marketing products, assessing their success, and charting new directions. And IBM might well sidestep cultural mismatch issues by laying off thousands of Sun employees.
The closest parallel from recent history is Hewlett-Packard buying Compaq–the deal that McNealy said in 2002 was “a slow-motion collision of two garbage trucks”. HP and Compaq had overlapping product lines, too, but fairly rapidly coalesced them, for example, by immediately displacing HP’s x86 server line with Compaq’s stronger line and by scrapping Compaq’s Tru64 Unix. What’s different about Sun and IBM is trying to figure out which Sun assets would emerge victorious over IBM’s.
The fact that a company as large and rich as Sun could be had for $4 billion or so must appeal to IBM, which doubtless expects to emerge from the current economic problems as a consolidator, not as the consolidated. But the acquisition would have to be justified not so much as fleshing out IBM’s already rich product portfolio, but rather on the basis of acquiring good engineers, a strong portfolio of intellectual property, a reasonable developer community, and one less competitor in the market.
This article was first published as a blog post on CNET News.