Oracle Pulls Commercial Support for GlassFish App Server
Oracle says it will stop providing commercial support for new versions of the GlassFish application server, though it will continue to deliver updates for the open-source edition.
Version 4.x of the app server will be the first released with no commercial Java EE 7 support; the 4.1 release of GlassFish Server Open Source Edition, announced at this year's JavaOne, is still on track for a 2014 release, the company says. That open-source edition will continue to be the reference implementation for Java EE.
Going forward, commercial Java EE 7 support will be provided from Oracle's WebLogic Server. And Oracle GlassFish Server 2.1.x and 3.1.x commercial customers will continue to be supported, according to the Oracle Lifetime Support Policy.
The company announced the plan this week in a blog post. Oracle wasn't taking calls about the announcement, but directed reporters to comments posted by Bruno Borges, an Oracle Java jock acting as something of a spokesperson on this news. In his blog, Borges clarified the details of the announcement and offered reassurances that "the GlassFish Community will remain strong towards the future of Java EE."
Few were surprised by the news. For some, it was a further realization of the fears the community expressed when Oracle bought Sun four years ago that it would, sooner or later, stop supporting Sun's open-source programs. Oracle no longer supports two such programs: OpenSolaris and OpenOffice.
For others, it was simply a foreseeable business decision.
"It looks like a step of rationalizing the two commercial products," said Ovum analyst Michael Azzoff. "The open source GlassFish will continue to support the very latest innovations coming out of Java, while WebLogic goes forward as the commercial offering with stability the No.1 concern, and a slower change cycle."
Referring to the OpenSolaris and OpenOffice decisions, Azzoff added, "These are commercial decisions and we know Oracle is more careful about the balance sheet than Sun was."
Oracle is already pushing GlassFish users to WegLogic, which the company has billed as the corner stone of its Cloud Application Foundation and a core component of its Oracle Fusion Middleware product line. Oracle WebLogic Server 12c was the first major update of the app server since the WebLogic 11g, released in 2009.
And that move won't be a cheap one, noted Forrester analyst John R. Rymer. "If you want supported Java EE, you must pony up for [WebLogic Server]," he said. "Folks who prefer GlassFish as their Oracle-supported production Java server now are looking at a big rise in costs. WebLogic Server is 5X the cost of GlassFish Server at list."
Borges argued that it's not necessarily true that WebLogic costs more than GlassFish. He pointed readers to the company's current price list (PDF here), and per-socket licensing language that allows licensees to count each CPU in a multi-chip module as a single occupied socket. "If you do the math," he wrote, "you will realize that WebLogic SE can actually be significantly more cost effective than [GlassFish Server Open Source Edition]…"
And yet it's hard to argue that moving customers to WebLogic won't add to Oracle's revenue stream, Rymer said. "Oracle is a profit-making enterprise," he said. "Sun was not."
Rymer said he expects fewer people to use GlassFish in production in the future, but he doubts that many organizations will actually be affected by this decision. "We don't have precise data on the size of the GlassFish community that buys Oracle support," he said, "but it seems small to me. Probably very small."
One reason for Oracle's decision to stop commercial support of GlassFish could well have been a lack of support from the GlassFish community itself, said Mike Milinkovich, executive director of the Eclipse Foundation.
"The only company that was putting any real investment in GlassFish was Oracle," Milinkovich said. "Nobody else was really stepping up to the plate to help. If you never contributed anything to it, you can't complain when something like this happens."
Milinkovich pointed to the latest blog post from Tomitribe, a Santa Monica, Calif.-based provider of Apache TomEE service and support, which also lays some of the blame for Oracle's decision on the GlassFish community. The blog compared Oracle's move to IBM's decision to discontinue commercial support for IBM WebSphere Community Edition last year.
"While it's tempting to wag our fingers at these two giants for their decisions to divest in any commercial interests in open source Java EE," the blog argues, "we might want to save a few wags for ourselves."
"There are a lot of people who are trying to create this narrative that Oracle is anti-open-source," Milinkovich said. "But I don't think that's true. Certainly within the Eclipse community, Oracle has had the Eclipselink project, which is the reference implementation for JPA, long before the Sun acquisition. And their investment in the project has continued apace. Oracle looks at open source as part of its business strategy, not as a let's-do-something-good-for-the-community strategy. And I have to point out, they're still here and Sun isn't, so it's hard for me to fault them. Altruism alone isn't enough to sustain an open source community."
If Oracle plans to keep GlassFish as the reference implementation for Java EE, then it can't be said accurately that the company has abandoned the app server. Milinkovich fully expects it to continue to serve as the reference implementation for the foreseeable future, which will require some contributions from Oracle.
"Everyone in the Java community has come to the realization that having open-source reference implementations is the best approach," he said. "I expect Oracle to put enough investment into GlassFish to keep it as the implementation of the current specs. The angst that you're hearing out there is because, even though it's a reference implementation, without real commercial support, GlassFish could become a toy, not something you would actually deploy into a running enterprise system. I think that concern is valid, and I would not be surprised if, over time, that came to pass."
Speculations about the future of GlassFish should also take into account the argument among some industry watchers that interest in application servers in general is waning. Redmonk analyst Stephen O'Grady points to a report by colleague Donnie Berkholz ( ("Interest Withering in Java Application Servers"), which argues that app servers are being "eliminated, component by component." Berkholz also noted that Oracle's acquisition of Sun "essentially put the open-source GlassFish into stasis, resulting in a downward spiral," adding "…GlassFish is well past its prime, post-Oracle."
It's also worth asking whether leaving GlassFish more or less on its own as an open source project is really such a bad thing, O'Grady said.
"The future of GlassFish will now depend largely on the size of the community and its interest level," he said. "The good news is that, like all orphaned open source projects, it has the opportunity to live on if the community is sizable enough and decides to support it, representing a large enough interest to attract the kinds of commercial support that are a requisite in many business cases. The bad news is that open source is no panacea; the project will have to live or die on its merits and market interest."