SailFin V2 released!
SailFin V2 (Sun GlassFish Communications Server 2.0) is released today. See Srikanth's announcement and Prasad's roadmap entry. A number of new features have been added since we released SailFin V1 in January. These include high availability, rolling upgrade, flexible network topology, better over load protection, improved diagnosability, Java based DCR files for the load balancer etc. Diameter support is also available.
The key enhancement in high availability is a new replica selection algorithm for the session replication feature as explained by Sreedhar. It leverages the consistent hashing algorithm the converged load balancer uses. This eases the load distribution after a failure and makes the cluster much more scalable than SailFin v1. Converged Load Balancer itself has been enhanced to support writing DCR rules in Java.
Rolling upgrade capability leverages the session replication. That means, when the cluster instance is restarted after upgrading the HW or SW, the sessions will be restored by retrieving the session replica that exist in the cluster. Bhavani explains all those here along with an example that showcases session replication and rolling upgrade. If you are trying out HA in sailfin, take a look at these tips and tricks.
Flexible network topology allow better usage of multihoming features in the platform. Here is a picture that shows one such deployment architecture. Ramesh's blog here is a very good source of information.
For more information on SailFin V2, please browse sailfin entries in the aquarium blog. Download and try V2 when you get some time. You may also try out the amazon images as explained by Sreeram. Send any feedback you may have to users alias.