Sun As An Independent Business Unit Of IBM

If IBM acquires Sun, and that prospect becomes increasingly likely with each day that passes without a denial, here's one thing it should consider doing: it should keep Sun intact and operate it as an independent business unit, much like EMC did with VMware.I'm not the author of this idea. I first heard it in a conversation with Frank Gillett at Forrester Research, but it's a good one. EMC could have made VMware's virtualization a feature on its storage systems, something like Microsoft makes Hyper-V a feature on the operating system. It didn't. As VMware grew, it encouraged the notion that it was an independent company by making 10% of it or so public. Today, instead of virtualization being a feature of EMC's storage, it's promising to take over the data center, and VMware is still leading the charge.

There are good reasons for the rest of us to have a stake in IBM acquiring Sun. Some of them are ably formulated by Bill Snyder in his column, Why We Should All hope IBM buys Sun, today. I can't vouch for all the reasons, and there's always the dangers implicit in what would amount to IBM's resulting two-thirds market share of the Unix market. But part of that market would be Solaris, or open source Unix. I don't think IBM would try to put the Solaris genie back in the bottle. I think it would try to wean Solaris users off UltraSparc and onto its Power-based servers.

But what I'd like to see is the Sun impulse to do things harnessed to the IBM capacity for restraint. I think that's a much more favorable outcome than Sun failing outright, which might be in the cards if the recovery doesn't arrive soon, or, say, being acquired by Microsoft when it's on the brink of failure.

IBM would pick up tangible assets, but the intangible involved is Sun's propensity to do things. Sun is a RISC chip architecture company, it's a CISC chip server designer; it's an operating system company, a storage company, a proprietary software company and an open source software company. Its software, in addition to Solaris for both Unix and x86 servers, includes application servers and other middleware, databases, identity management, two different virtualization hypervisors and, horrors, the newest scripting language on the planet, Javafx, as if programmers have been yearning for yet another scripting language.

Sun once entered into an agreement with AOL to take on Netscape software and manage its future generations. But Sun has not developed the disciplines of a mature software company. It couldn't manage the bewildering number of choices presented by Netscape software because Sun doesn't make choices. It's a systems company, which means it makes everything that systems need and they need everything Sun decides to do. The Sun-Netscape linkage was a train wreck in slow motion, and scarcely anything of value survived.

If Sun is management challenged on the software front, it's not necessarily true that IBM operates in the public interest. I think it operates in its own interest, as it should, and it knows its own interest very well. It survives through generations of different management, while other companies succeed on the visionary strengths of one management team, then wither and disappear. Sun began its trek to success with Scott McNealy and it started heading for the exits under the same leadership.

The acquisition of Sun by IBM remains an unlikely prospect for success. The Sun impulse to create would probably be overwhelmed by the disciplines and procedures of Big Blue; at the very least culture conflict would be sure to occur. But if IBM could hone Sun's priorities, focus them on goals with long term prospects and curb its impulse to compete on every front, then let it pursue those goals, the result might be surprising.

Sun as an independent IBM business unit might produce a starter kit of open source software for small and medium business. It might introduce these new customers to cloud computing and provide them with services that let them grow their business before expanding their data centers.

Rebecca Wetteman at Nucleus Research warns that IBM's previous acquisitions, such as Lotus Development and Rational Software, didn't lead to radical improvements in what those companies could do. In the case of Lotus, it may have hampered it, although Lotus had made its own set of decisions that started it toward the exits before IBM showed up on the scene.

Sun customers might be better off if IBM acquired Sun and simply incorporated those parts of the Sun product line that fit into its existing portfolio, such as Identity Management, and discarded the rest. But Sun as a trimmed down independent IBM business unit, with IBM deciding on the priorities that got funded, remains an intriguing prospect.