What a Sun-IBM merger could mean for Java

The acquisition of Sun by IBM raises several interesting possibilities, especially in regard to the wide array of Sun software that is likely to be affected. Here are some thoughts on the effects of the deal on key technologies.

Java. Several pundits have speculated that if IBM controlled Java, the language and platform would be even more open source than it is, and that problems such as product conformance tests for OSS versions of Java would go away. This is probably correct. However, the one thing that will not happen is IBM ceding control of Java to a third party. In other words, while the JCP might be dismantled (though I doubt it), some system will remain in place by which IBM can pursue its agenda in Java.

The reasons for this are woven into IBM’s genetic material. Big Blue has always had a strong preference for OSes and execution environments that it can control. This preference took what the company viewed as a very bad turn when IBM teamed up early with Microsoft on the PC. All went well until the infamous spat that led to the Windows-OS/2 split.

When that occurred, according to what IBM insiders told me years ago, the company recognized that it could never again put the fate of its operating systems in the hands of a third-party vendor. For that reason, IBM was attracted to Java for its enterprise apps. It could run them on its own and just about any other operating systems. Because of this, it enthusiastically endorsed Java.

However, IBM soon realized that Sun would only let it in so far. Sun’s wariness of IBM has been a longstanding sore spot for Big Blue. So, having acquired Java in this transaction, IBM is hardly going to turn over control to another third party. Expect the direction of Java to be influenced by IBM for the foreseeable future.

Enterprise Java. WebSphere is a well-established Java EE server, while Sun’s various EE server initiatives have gone nowhere. I expect the Sun EE offerings will fade away; the open-source software products could live on, but I doubt IBM will give them enthusiastic support.

Java Tools. IBM has already developed a solid IDE that it turned over to the Eclipse Foundation. I can see no scenario in which it would adopt or extend NetBeans. I suspect NetBeans will be cut loose as an open-source project to live or die by how much its existing community supports it. A similar fate might well befall language projects like JRuby.

Even though it only takes a couple of engineers to keep a language moving forward, I am not sure IBM will see much value in this. And IBM’s record in new language development is not great. You might recall that the first major scripting language for Java was IBM’s NetRexx, an excellent product that I have discussed several times in my column. But IBM put little effort into marketing it, never open-sourced it, and eventually reassigned the principal engineer to other projects.

Databases. MySQL is the obvious prize here. Rebranded as part of the DB2 product line, it could easily be Big Blue’s ticket into the lower end of the market, that is, the sub-enterprise and SMB space.

Right now, IBM’s free version of DB2 is akin to the free version of Oracle: a scalability-limited placeholder for small businesses that might want to try out the databases. MySQL, by comparison, has wide industry support and no artificial limits on database size or number of processors, so instead of being a push tool like the free DBMS products, it can become a pull product that encourages users to step up to paid versions of DB2. Pulling that off will require careful marketing by IBM, a company not known for its prowess in that area.

Operating Systems. Solaris is a far superior version of Unix than AIX. In the Unix community, AIX has always been viewed as a technologically weak variant. I expect that IBM will either migrate AIX customers to a relabeled version of Solaris, or it will migrate some superior features of Solaris (such as ZFS) to AIX, thereby significantly bolstering its capabilities. Some of these features are likely to appear in IBM’s Linux as well.

Beyond the software components are the hardware aspects of this deal: IBM will surely close down the SPARC processor line over several years and move the SPARC engineers to the Power and PowerPC processor design team. SPARC Niagara was an innovative product and IBM will surely want to hang on to those designers. Sun’s storage business will be something IBM either sells off or quietly folds into its enterprise storage product line.

Overall, in my opinion, for the few billion dollars this deal costs Big Blue, this acquisition is a stellar bargain.