By alexismp on Mar 04, 2011
People usually expect performance improvements from "dot 1" releases and that's a fair aspiration for GlassFish 3.1. Performance can obviously mean different things to different people but we believe there's good news for most of you with this release of GlassFish 3.1.
Application server veteran and benchmark guru Scott Oaks discusses the performance enhancements in this release from both the developer and high-availability (HA) perspectives which show each improvements in the double digits over the 3.0 release.
Scott's blog post covers improved startup, deploy and redeploy times and the benefits of the modular architecture introduced with GlassFish V3 because "in the end, pure startup isn't what is important -- what's important is how quickly you can get all of your work done".
Because this is not a new area for GlassFish, the HA part also got its share of performance improvements building on the work delivered in previous incarnations and derivatives of the product. In particular both the GlassFish underlying Grizzly-based nio framework and the in-memory replication have seen session serialization and general implementation improvements enhancing HA's both full session and modified-attribute replication scopes.
When it comes to performance, the Performance Tuning Guide is probably a good place to start.