23 Oct 2012 , tagged: OSGI, Java

OSGI Bootstrapping

I’ve been working with OSGI based technology for a while now and it’s a great piece of technology. However, so far I’ve only worked with OSGI technology that runs in existing instances of the framework. A running framework is easy to deal with, but how do you get to this point? I spent some time today playing with that, and it took me a while to put it all together. So naturally, I’ll have to record my findings here.

Section 4.2.1 of the OSGI specification (version 4.3.0) explains the process in minute detail. Here’s a short rundown of how it works:

Every framework implementation must provide a org.osgi.framework.launch.FrameworkFactory that produces instances of the framework. There’s several different ways to retrieve the actual class to be used. It could be hardcoded, retrieved via command line parameter or via Java service loader
Once you have your FrameworkFactory instance, it’s as easy as calling frameworkFactory.newFramework(Map<String, String> configuration). The configuration Map is a bunch of key-value pairs with configuration parameters for the framework (and the implementation). From what I gather, all the OSGI specific keys are defined in org.osgi.framework.Constants. One example is the Constants.FRAMEWORK_STORAGE key, that defines where the framework is supposed to store its bundles.
The newFramework() method will create a instance of org.osgi.framework.launch.Framework, representing the entire OSGI runtime. You’ll have to initialize it by calling init() and then start it with start(). Once start() returns, your framework should be up and running.
In order to avoid the framework being shut down right away, it makes sense to call waitForStop(int timeout), a synchronuous method that will block until the framework is shut down.

That should be enough to get your framework started.