JBoss Data Grid
A good candidate to run atop Hadoop might be JBoss Data Grid 6.1, Red Hat's in-memory database. In a major update released in April, the massively scalable database received significant new functionality for high availability and disaster recovery. Red Hat's database for large-scale enterprise applications now supports data-center replication across geographically dispersed clusters as well as the ability to perform rolling upgrades without interrupting service. A new nonblocking state transfer implements an all-new join-and-leave protocol for grid nodes that does not cause the system to pause, enabling databases to continue responding to requests while it adjusts for a missing or failed node.