New ActiveStor 5000 and ActiveStor 3000 Enable Corporations and Research Organizations to Accelerate Their Time to Results With a Unified Approach to HPC Storage: Panasas, Inc., the leader in clustered storage solutions for the High Performance Computing (HPC) market, today announced two new families of storage clusters, the Panasas ActiveStor 5000 and Panasas ActiveStor 3000. Both ActiveStor 5000 and ActiveStor 3000 include the new ActiveScale 3.0 operating environment and Panasas PanFS Parallel File System. With the introduction of the ActiveStor Storage Clusters, in conjunction with the ActiveScale 3.0 operating environment, Panasas now offers the most comprehensive portfolio of storage solutions for HPC applications. Panasas ActiveStor 5000 is targeted at HPC end users who desire a unified storage fabric and have a mix of batch applications that run on clustered Linux servers and interactive applications that run on workstations. The ActiveStor 5000 is offered in 5 Terabyte (TB) or 10 TB configurations scaling to 100 TBs on a single fully configured, pre-wired and tested Panasas rack system. By unifying storage for both types of HPC applications, organizations can speed the workflow across their product development and research environments. In addition, operating costs can be substantially lowered by minimizing duplication of data and dramatically simplifying storage management. Applications that can take advantage of the ActiveStor 5000 product line include seismic interpretation, financial risk analysis, bioinformatics, electronic design automation, computational fluid dynamics and special-effects animation creation. "Storage requirements typically vary across an organization depending on the HPC application, with interactive users demanding higher data availability, high I/O capability, and more random access, while batch processing users are more focused on extreme capacity and high sequential I/O performance," said Addison Snell, research director of High-Performance Computing at IDC. "Although previous Panasas products were frequently used to optimize data access for individual applications, Panasas is broadening its appeal to HPC cluster managers who are seeking those benefits across their end-to-end workflow." Included in the ActiveStor 5000 are capabilities that enable the new Panasas storage solution to address performance demands of a full HPC spectrum of applications, while delivering the highest capability in large scale data protection: -- Superior Performance -- With the new StorageBlade XC capability, the parallel random I/O performance allows substantially faster access to data in memory instead of having to access data from the disk. The StorageBlade XC increases the StorageBlade cache from 512MBs to 2GBs. This higher speed cache increases I/O performance by as much as 10 times with up to 20GB of cache per Panasas shelf, or up to 200GB per fully configured rack system. -- Capacity -- The new StorageBlade XC offers up to 1 TB of capacity per blade. This enables each ActiveStor 5000 shelf to support up to 10 TB, with a full rack system supporting up to 100 TBs in a single 42U enclosure. -- Data Protection and High Availability -- The new ActiveGuard product ensures non-stop processing and high availability to ensure critical uptime for a unified storage infrastructure. If a failure occurs, it is quickly resolved with the use of a second Gigabit Ethernet link for complete storage network failover, as well as a Panasas DirectorBlade metadata failover facility to safeguard the information about the data. -- Volume Snapshots -- The new ActiveImage product enables volume snapshots to ensure maximum data availability and accessibility. Over 32,000 images per storage cluster are supported. ActiveImage minimizes storage requirements by capturing only changes from the previous image. "The latest release of Panasas' ActiveScale operating environment and ActiveStor 5000 has enabled us to even further increase the productivity of our Linux cluster. Panasas continues to provide powerful solutions that address our demanding HPC cluster needs for ultra high performance, management, and reliability," said Wesley Emeneker, HPC research assistant at Arizona State University's Fulton High-Performance Initiative Center. "Our researchers depend on high bandwidth and data availability for simulations including weather forecasting, snapshots of the moon, and computational fluid dynamics. A large portion of this data is irreplaceable. Panasas' new products including enhanced data protection and high availability features allow our researchers continuous uptime and data security for all of their processing needs." The new ActiveStor 3000 is optimized for low-cost, high-capacity, batch-oriented Linux cluster applications such as seismic processing, EDA simulation, and large-scale modeling where superior scalability and low cost per gigabyte are the primary buying criteria. For even higher capacity, the addition of a new 1 TB StorageBlade delivers up to 10 TB of storage capacity per 4U shelf, and up to 100 TBs in a single rack system. In addition, multiple cabinets can be configured in a single storage cluster that can scale to petabytes of capacity to simplify and reduce installation time to as little as 15 minutes. Customers can create a pipeline of jobs to be processed to maximize the use of their compute clusters, improve IT productivity, and accelerate time to results. "Our mission at Panasas is to provide the highest performance and easiest to manage integrated storage solutions for HPC customers. We have proven our ability to meet the requirements of the most demanding Linux cluster users in the federal government, academia, and commercial enterprise segments," stated Victor Perez, president and CEO of Panasas. "These new products will bring the unsurpassed performance and manageability of Panasas ActiveScale storage solutions to a much broader set of customers -- those that require a unified storage infrastructure for all of their HPC applications." ActiveStor 5000 and ActiveStor 3000 will be released to new and existing customers in January 2007 and will be showcased at the upcoming SC06 Supercomputer Exhibit and Conference during the week of November 13, 2006.