Slashdot Mirror
Oracle Kills Commercial Support For GlassFish: Was It Inevitable?
An anonymous reader writes "Oracle acquired GlassFish when it acquired Sun Microsystems, and now — like OpenSolaris and OpenOffice — the company has announced it will no longer support a commercial version of the product. Mike Milinkovich, executive director of the Eclipse Foundation. said in an interview the decision wasn't exactly a surprise: "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." An update to the open source version is still planned for 2014." GlassFish is an open source application server.
What the hell is a glass fish?
I don't believe it. Right there in the summary.
From my experiance, Glass fish is ONLY used by people following the JEE tutorials from oracle (using netbeans too). It is not a competative-performant-scaleable JEE Application server.
You seem to have limited experience then. Glassfish is the reference implementation of Java EE standard and therefore it is used in JEE tutorials. BTW, IT IS used extensively in many enterprise application, including very demanding stuff like stock broking and trading (I have designed it for a large customer myself who serve more than million trades a day, so I can speak with some authority). This is a big news exactly for the same reason. There are many enterprise customers who paid money to get commercial support on Glassfish. Now those companies will either have to depend of the community for support or switch to other commercial options like WebLogic or WebSphere or JBoss EAP.
Here is a better answer (and it's not mine). The 0.1% figure is disingenuous. People normally don't user app servers like Glassfish as web servers. They usually serve back end web services (which doesn't really count as a web server in my book, despite the terminology used), and back end Java Enterprise services (like Enterprise Beans, and Persistence Layer objects). In my experience, while many use Tomcat for web services, it is a pain in the ass to use for any seriously large sized projects. And it is kludgey and tougher to configure unless you like playing with Apache style configuration files (meaning they are about as clear as Apache documentation). Glassfish is built with all the services required and integrated for doing most anything you need to do with an app server, no added packages needed. People who will tell you Glassfish isn't very good are also those who still think Netbeans is no good, when in fact it now eclipses Eclipse for just working without fucking around with adding plugins. And it works very well. Also Glassfish has built in facilities for horizontally scaling/high availability.
-- I ignore anonymous replies to my comments and postings.
Sounds like this is wonderful news for you guys. You both have clients that are loaded with money, and who desperately need Glassfish support for their production environment.
And now Oracle stops offering support? Dude, this is the best business opportunity you'll get in your life. Quit whatever you're doing and start offering Glassfish support yourself. If it's really that big a deal, companies will be all over you.
There's probably a lot of truth in what you say. Banks tend to be incredibly conservative in their upgrades. I know of a couple that are just finishing their migration from FreeBSD 2.x to FreeBSD 6.x. They didn't even manage to start the upgrade to 6.x until after it was no longer supported upstream, but they pay people to backport security fixes. I know of a couple of others that still do a lot of processing on VMS on VAX, although they do have some Alpha and Itanium boxes. Outside of HFT, performance doesn't matter that much to them, but they really don't like surprises. This is why they tend to be a lot more positive about open source than you might expect from such a conservative industry: they like the idea that you can keep running an ancient platform long after the original vendor goes out of business. Transaction processing volumes grow a lot more slowly than Moore's Law, and unless they need new features they'd much rather keep using the system that they know works with occasional bug fixes when it doesn't than have to switch to something newer.
I am TheRaven on Soylent News