IBM-Sun Rivalry May Finally Reach An Accord 2IBM-Sun Rivalry May Finally Reach An Accord 2

If IBM succeeds in acquiring Sun Microsystems, it will lay to rest one of the bitterest rivalries between erstwhile allies that the industry has seen in a long time.

Insiders close to both companies suggested this week that IBM is in talks to buy Sun for $6.5 billion. Both companies have officially denied the reports as speculation.

One of the main reasons IBM would be willing to pay a premium for Sun is that Sun owns Java, and Java IP is worth billions in future revenue to IBM. IBM's software lines, such as WebSphere and Rational, are heavily Java-based. Yet IBM doesn't pay Sun annually for a Java license -- or at least it didn't in the late 1990s when a dispute broke out between Sun and IBM on that point. Sun said all commercial Java vendors needed a license; IBM responded that its early Java agreements with Sun covered its existing uses of the language.

It was a sign of how much IBM's early commitment meant to Sun that IBM could make such a rejoinder. Sun grumbled and rattled the license saber occasionally but IBM bided its time and kept its lawyers primed. It was one of several sources of tension between the two companies most responsible for establishing Java. Now it's the language believed to be taught in more computer science classes than any other.

Another tension was over Java tools. Sun found itself in conflict with other members of the Java coalition as Sun gained early traction with products incorporating the latest Java -- developed inside its halls. Other tool vendors, who also helped get Java established, objected. Sun backed away from offering Java tools and established a quasi-open Java Community Process, where control over additions to Java would be shared in working groups and review committees.

When it re-entered the tools business, IBM took a programmer's workbench that it used internally and in 2001 made it open source. The workbench allowed different tools to plug into a shared environment and exchange files. With Eclipse, so named because it was meant to eclipse Sun's new tools drive, IBM established a rallying point for Java tool vendors in their competition with Microsoft. So many different toolmakers flocked to Eclipse that Sun found its huge investment in tools not yielding the return that it had hoped. Eclipse's success may have been Java's success, but Sun's tools initiative suffered and Sun never joined the group.

Now, if IBM acquires Sun, one of the decisions it will have to make is whether to add Sun's NetBeans tools to its product line. Chances are, it would leave it independent as an open source starter kit that would bring more developers to Java, and, eventually, to Rational. Another software issue would be what to do with Solaris. Solaris would bring into the IBM camp a host of financial services and telecom companies that Sun succeeded in capturing. And there's no technical reason why Solaris, if it can be ported to the x86 instruction set, couldn't be ported to IBM's Power architecture. Power chips need more customers, and Sun's Sparc processor customer base would be one place to find them.

IBM would probably be able to keep Solaris as open source code without offending members of the Solaris community. It's honed its skills at dealing with open source communities to keep conflicts to a minimum.

But IBM's oft-praised early support of open source code sometimes overlooks the fact that it fits open source into its long-term business plans. It has selectively supported projects, and oftentimes its support was based on not only respect for the timeliness of an emerging piece of open source code but also the fact that it might undermine a competitor. Its early support for Linux helped Linux nibble away at Sun's customer base and undermined Sun as a Unix competitor.

Coming out of manufacturing workstations, Sun occupied the low end of the Unix market and broadened it aggressively with its Solaris operating system. Linux started competing directly for the same users, as x86 PCs took on more workstation characteristics. For many years, former Sun CEO Scott McNealy couldn't refer to Linux without making a derogatory or at least conflicted remark. It took a change in CEOs to the more open source-minded Jonathan Schwartz to make peace with Linux and adopt open source code as a long-term trend that Sun might profit from as well. If IBM acquires Sun, observers may one day conclude that the transition occurred too late.

One thing is clear: If IBM buys Sun, Sun's somewhat-scattergun open source strategy will be tied more closely to a specific business plan. IBM supported Apache and bundled it with its proprietary products when it saw it might best Microsoft's IIS. Apache did. But when JBoss got a little too big for comfort as a free Java application server, IBM breathed new life into Geronimo, a Java application server project founded by JBoss breakaway developers. Today, Geronimo is offered as WebSphere Express, a free product meant to attract independent developers into the IBM software stack.

If Sun's open source strategy isn't working under downturn conditions, why is IBM willing to pay the reported $6.5 billion, or 100% more than its closing price before word of negotiations leaked out?

The stock market had been valuing Sun at a total capitalization roughly equal to the amount of cash and cash assets that it had on hand. That was the market's vote of no confidence in Sun's ability to make it through a harsh recession. IBM is nevertheless willing to pay more than what Sun is theoretically worth because without speculative support, Sun's stock could sink even lower. What if it uses up its cash reserves and starts reporting a string of losing quarters again?

At the rate its stock was going down, Sun was going to be available at fire-sale prices if somebody didn't do something.

What if Microsoft showed up at the fire sale and Java IP disappeared at a bargain price behind the walls in Redmond, Wash.? Perhaps just as bad, what if the company most adept at competing with IBM in consulting services, Hewlett-Packard with its EDS acquisition, showed up and acquired Java IP and Sun's software? Would HP CEO Marc Hurd and EDS know how to play that hand?

IBM doesn't want to find out. It doesn't mind paying a little more to keep these competitors out of the bidding, if it gets a lockdown of Java technology.

information Analytics has published an independent analysis of the challenges around virtualization management. Download the report here (registration required).