IBRIX Demonstrates Record I/O Bandwidth with an Embedded Storage Computational Cluster at Yale University; Breakthrough Performance Tests Running on New Cluster Architecture.

BILLERICA, Mass. -- IBRIX(R) Inc., the leader in scalable file serving solutions, today announced that a 57-node computational cluster at Yale University has demonstrated 10 gigabytes per second of aggregate I/O throughput from a single directory in a 14 terabyte single namespace file system utilizing IBRIX Fusion(TM). The new cluster architecture, comprised of a computational cluster with a single namespace IBRIX parallel file system created from storage embedded in the compute nodes, makes cluster computing an order-of-magnitude or higher more efficient and economical than with current architectures.

"Advances in computing to date have emphasized the role of increased CPU power but developments have lagged in deploying, accessing, searching, and mining large data sets ranging from terabytes to petabytes and beyond," said Dr. Steven Orszag, Percy F. Smith Professor of Mathematics and Computer Science, Yale University. "The impact of our development on scalable computing is significant. Our cluster opens a new era in data-intensive computing in which scalability limits are largely removed. This is done by bringing computational processing to the data rather than the other way around. At Yale, our applications range from studies of turbulent flows to exploring the geometry of massive datasets in huge numbers of dimensions. These applications and others, such as bioinformatics and complex image processing, can be done at significant cost savings compared to traditional clusters."

In traditional clusters, the data storage is centralized and separated from the computational parts of the computer. Enabled by the IBRIX Fusion parallel file system with its patented, segmented architecture, a single image file system is created from embedded data pools that are directly attached to the computational elements, thereby closing the gap between computer capability and storage access and throughput. With this distributed, embedded approach to storage architecture, extremely high application performance is realized at substantially lower cost.

"Efficient knowledge extraction from massive data inputs requires a distributed data storage system as demonstrated by our cluster at Yale," said Dr. Ronald R. Coifman, Phillips Professor of Mathematics and Computer Science, Yale University, and a winner of the National Medal of Science. "It is precisely this capability that enables a self-organizing massive organic memory in which items linked in huge numbers of ways encapsulate knowledge."

"The tremendous flexibility and scalability of the IBRIX Segmented File System(TM) architecture is enabling a radically new approach to high performance computing that will reduce the cost of scale-out compute clusters substantially," said Shaji John, Chairman and CEO, IBRIX. "This innovation will further the expansion of scale-out computing into new realms of scientific and commercial computing applications."

IBRIX Fusion(TM) Scalable File Serving Solution

IBRIX Fusion is a software-only, highly-available, scalable file serving solution that allows enterprise system administrators to build file systems that can scale to up to 16 petabytes of capacity in a single namespace, and provide up to one terabyte per second of aggregate I/O throughput performance independently and non-disruptively. Equally applicable for both high-performance/high-throughput applications as well as metadata-intensive, small block enterprise applications, the multipurpose IBRIX Fusion file serving solution delivers I/O performance and scalability that is greater than that achieved by other file serving solutions. IBRIX Fusion is hardware, network, and protocol independent and can be deployed as either a host-installed cluster file system solution or exported over network file system (NFS), common internet file system (CIFS), or other industry standard protocols as a scalable network attached storage (NAS) solution. IBRIX Fusion enables a single image file system to be created from both networked and direct attached storage.

About IBRIX, Inc.

IBRIX(R) delivers increased application performance to enterprise computing environments utilizing a unique, software-based parallel file serving solution. The IBRIX Fusion(TM) software suite, comprised of a single namespace parallel file system, a scalable volume manager, high availability features, and a comprehensive management interface, is the most scalable, economical, multi-purpose file serving solution available for cluster computing, grid computing, and enterprise computing environments. Based on IBRIX's patented Segmented File System(TM), IBRIX Fusion increases application performance, simplifies administration, and reduces total cost of ownership. The IBRIX Fusion software suite is available through leading OEMs. For more information, visit www.ibrix.com.

IBRIX, IBRIX Fusion, and IBRIX Segmented File System are trademarks of IBRIX, Inc. All other brands, products, or service names may be trademarks or service marks of the companies with which they are associated.