Terracotta’s Naturally Clustered Java lets developers view   applications as if they were running in a single Virtual Machine (VM), and enables runtime clustering capabilities that enhance application scalability and availability, the San Francisco-based company says.

Clustering Java, using the Terracotta Virtualization Server (TVS),  increases the stability and fault tolerance of Java applications while simultaneously reducing ongoing development and operating  costs. TVS Eliminates long-standing tradeoffs developers have been forced to make between availability and performance in Java applications.

With TVS, availability and performance come together in a “plug-configure-play” infrastructure software package. This means that there are no code changes, no APIs, and no altering of application semantics. TVS provides complete data access visibility and control  to development or operations staff at runtime. Developers can use natural Java with no need to learn, use, or maintain proprietary API’s.

TVS has two primary components that together meet a diverse, but often concurrent, set of data access and consumption needs; Distributed Shared Objects (DSO) and High-Availability-JDBC (HA-JDBC).

DSO is a Terracotta Virtualization Server component that provides the capability to share object state across Java Virtual Machines (JVMs). DSO eliminates the need to maintain state in a transactional database, as well as the need to rely on inter-JVM communication mechanisms like JMS, RMI, Jcache, serialization, etc.

DSO’s features:
• Inter-VM object sharing on opt- in/opt-out basis
• Flexible locking configurable at runtime
• Virtual heaps for multi-gigabyte object graphs
• Scalable multi-level clustering infrastructure
• Distributed method invocation
• Distributed wait and notify

DSO’s benefits:
• Shared data without database load or abuse
• Transparent recovery from application instance failures
• Clustering as an API-free infrastructureservice
• Reduced development, testing, and maintenance costs
• Faster application development and scaleout
• Fine-grain operations performance visibility and control

HA-JDBC is an API-free Terracotta Virtualization Servercomponent that 
provides an intelligent two level JDBC caching mechanism across Java 
Virtual Machines (JVMs) in a “share everything” mode. HA-JDBC 
significantly reduces the need to touch the database directly.

HA-JDBC’s features:
• Caching allows read-through of DB queries, write-through for user 
changes
• Scalable multi-layer infrastructure
• Cache invalidation directly through our database Listener
• Opt-in/Opt-out at query level
• “Pre-warmer”warms all caches in batch mode
• Cache Pinning returns current cache data when DB is unavailable

HA-JDBC’s benefits:
• High-performance data access
• Decreased DB uptime dependence
• Fine-grain production-time visibility and control over performance 
of DB queries
• Reduced database load

Specifications: The Terracotta Server is a pure Java solution that  should run on any platform that supportsJRE 1.4 or above and is supported on Sun’s HotSpot VM on Linux,SPARC Solaris,and Windows XP/WindowsServer 2003.The Terracotta Listener supports Oracle 9i for SPARC Solaris,and Linux Kernel version 2.4 or higher.

It is available for downloading at:
http://www.terracottatech.com/company/registerme02.html