Systems/Enterprise:

PAREMUS SELECTS BLITZ JavaSpaces FOR INFINIFLOW ENTERPRISE GRID

Paremus announced its sponsorship for the Blitz open source project and commercial support for Enterprise Blitz, in order to advance the deployment of Jini and JavaSpaces technology with Infiniflow in mission critical business environments.

Developed by distributed systems specialist, software engineering consultant Dan Creswell, Blitz is a high throughput, standards based, JavaSpace implementation. Blitz also provides a rich set of features to make the development and deployment of JavaSpaces technology easier and more efficient.

JavaSpaces, a component of the Jini framework developed by Sun Microsystems, is a distributed, associative, network-accessible, shared memory system implemented as a Jini service. At the highest level, a JavaSpace provides synchronized, transactional, read, write, and associative search access to objects in a persistent shared networked memory. Clients to a JavaSpace can use it to share information, including elements of an application that can be solved in a distributed and parallel fashion within compute, data and transactional Grids.

"Widely deployed across a number of industries, Blitz is the leading open source implementation of Sun's JavaSpaces technology, and is a core component of Infiniflow architecture," said Richard Nicholson, Paremus CTO. "By providing commercial support for open source Blitz, we are allowing our clients in tier-one financial institutions to have confidence leveraging Blitz at the core of their mission critical environments, and confidence in JavaSpace technology in general, by increasing competition and choice in commercial supported versions of the technology."

The next generation of distributed enterprise compute solutions must be able to dynamically reconfigure and reroute information and processes, so, as a part of recent development work, Paremus evaluated the suitability of numerous middleware solutions. Blitz was selected, not just for its very high transactional object throughput and server decoupling characteristics, but also for its flexible deployment and re-configuration capabilities. This set of features meant that Blitz was the optimum persistence-messaging substrate layer for the Infiniflow architecture.

Infiniflow's adaptive and self-healing behaviors combine information provided by the Blitz performance interfaces with raw machine and network statistics in order to dynamically retune and optimize resource usage, even when large parts of the compute infrastructure have failed outright.

"Testing carried out in conjunction with Paremus and Sun has allowed us to quickly ensure Blitz achieves the stringent requirements of the Paremus enterprise Grid architecture," said Dan Creswell, Blitz founder. "Where distributed debugging has historically been a major challenge, Blitz's remote debugging interfaces have allowed us to quickly pinpoint and improve performance issues across arrays of blade and rack servers."

The versions of Blitz supported by Paremus are the same as those used in Infiniflow product releases. As such, these versions of Blitz will have been subjected to extensive in-house stress testing by the Infiniflow engineering team. Clients that subscribe to full commercial support will also have access to Paremus' distributed application design services as well as custom Blitz application tuning and benchmarking services.

"Our intentions are to allow clients to tap into our leading distributed systems design skills within the Infiniflow engineering team, and provide client access to Paremus and partners Infiniflow stress test environments," said Mike Francis, business development manager at Paremus. "We are already working with a number of financial institutions to help them exploit the utility computing model and provide unprecedented business agility, service availability and cost effectiveness. These Blitz centric services will help accelerate such client initiatives."

"I'm delighted that Paremus has chosen to adopt Blitz as an essential part of the Infiniflow architecture," said Creswell. "Paremus is a first rate consulting and support organisation and shares my vision of distributed computing, with Jini and JavaSpaces at the heart of an autonomic compute fabric."