|
INTEGRATION WATCH:
Open-Source Enterprise: A Mixed Prognosis
By Andrew Binstock
 November 15, 2004 — This fall, JBoss released version 4.0 of its Java application server—the first open-source product to receive the official sanction of J2EE compliance from Sun.
This certification is a big deal. Some studies show that JBoss is the fastest-growing Java platform among developers. And in this niche, its place is secure because JBoss integrates two technologies that developers know and like: the Tomcat container and the Hibernate persistence engine. To these, JBoss adds its own caching engine. I previously have discussed the importance of excellent caching mechanisms in enterprise Java (“Caching In on Java Caches,” Aug. 15, page 29), and they are included in JBoss’ solution.
To make a go of this now-ready-for-the-enterprise product, JBoss will have to button down a bit. Right now, the product is leased as an embedded platform to ISVs and used as a development platform from which code is then migrated (presumably to BEA’s WebLogic or IBM’s WebSphere). These positions hardly have the makings of an enterprise product.
However, I predict good things for JBoss because it has the key ingredients of a successful company: a good product, a good executive team, an ability to generate attention, and a warrior’s relish for the fight. With the accelerating acceptance of open source in the enterprise and the J2EE diploma, JBoss could—with some luck—snag the bronze medal (gold and silver will be IBM’s and BEA’s for a long time). In the process, JBoss will, I believe, push aside Oracle, Sun and the smaller niche competitors (Macromedia, Persistence and Sybase).
If this path to success seems too easy, that’s because it rides along an odd market fault line: Vendors of Java application servers went into a big snooze once IBM caught up with BEA. Unwilling to invest into grabbing market share for permanent third place, they stopped pushing and just let the market grow as its internal forces would dictate—an almost perfect scenario for a motivated player with a good product. Enter JBoss.
Not all enterprise markets will be won over so easily by spirited open-source vendors. In particular, the database market will remain, in my view, very much the way it has been during the past few years: IBM and Oracle running neck and neck, Microsoft’s SQL Server in third, Sybase far behind the leaders but competitive on the strength of its enduring popularity in financial services, and then a bunch of small players staggering after the leaders.
The biggest difference from the app-server market is that the big three vendors are all seriously committed, vigilant and ready to meet all challengers. Oracle, of course, is the poster boy for pugnacious defense of its database turf. Microsoft is no shrinking violet. And IBM overtook Oracle—granted it was on the basis of the Informix acquisition—but after catching Oracle in databases and BEA in app servers, no one can say IBM does not respond to market challenges.
Bottom line: MySQL, which undeniably has come the furthest in marketing an open-source database, is not going to break into the glass house anytime soon.
In this context, IBM plays a most interesting role. Its reputation for supporting and embracing open source is unassailable due to truly extraordinary contributions to the movement. Yet IBM has no problem disrupting any pursuing open-source upstart by dropping marbles from the back of the truck.
Just this summer, IBM turned over its embedded Java database, Cloudscape, to the Apache Software Foundation. (By the by, it is a terrific product that’s completely free. It has the full complement of high-end features, including stored procedures, which MySQL still lacks. Cloudscape can run in a server configuration as well and hold databases limited only by disk size.)
The effect of Cloudscape and any other midrange database product IBM might care to open up is to disrupt poaching by upstarts. Sybase recently joined the party with the release of a free version of its flagship database on Linux (with some limitations that keep it at the low end). If open-source or free products are not sufficient, the big three will lower prices in the midtier and live off the enterprise packages, whose sales will be tied into other products (think PeopleSoft for Oracle and a slew of apps for IBM). Microsoft will be the least capable of playing this game and could be somewhat pressured.
JBoss and MySQL are not comparable in their prospects: JBoss will undoubtedly find its way to the big game, while open-source databases—all of them—will be shoehorned between enterprise packages and entry-level products and constantly pressured from both ends.
Andrew Binstock is the principal analyst at Pacific Data Works LLC.
Andrew Binstock is the principal analyst at Pacific Data Works LLC
|
|
Advertisement
Click here for a complete listing of Integration Watch Columns
Click here to see a complete Column Archive. |
Back to Top
|