This is an implementation of a replicated hashtable in Java that will autodiscover other nodes on the same machine/network subnet via multicast and replicate data to them.
In use, it's completely transparent, requires zero configuration and acts as a drop-in replacement for java.util.Hashtable. it's great for adding failover to your servlet applications (and is used by OOWeb for storing clustered session information)
It comes as a single class that you can add into your project - no bulky libraries required.
Requires any free Java virtual machine.