EARTH SCIENCES
DataDirect Networks' S2A9550 Storage Solutions Power TeraGrid
DataDirect Networks Delivers 32.5 GB/s With 1.46 PB to IU's Big Red Supercomputer and Data Capacitor High Speed Storage Repository - Ranked as One of the World's Fastest Supercomputers: DataDirect Networks announced that it has shipped 14 of its latest 7th generation S2A9550 high performance storage cluster solutions, totaling 1.46 PB, to Indiana University's (IU's) Big Red supercomputer and Data Capacitor High-Speed Storage Repository.
The S2A9550 storage cluster solution is based on DataDirect's industry proven S2A technology which fully integrates a parallel, non-blocking architecture with SAN functionality in one simple, easy-to-use intelligent device.
Big Red is part of the National Science Foundation (NSF) and is an integral part of NSF's plan to create an advanced national cyberinfrastructure, called TeraGrid. Big Red is one of the fastest supercomputers owned and operated by a U.S. university, and is the largest IBM e1350 system in the world.
Data Capacitor is a storage system innovation developed by IU to help scientists and researchers deal with the massive amounts of data being generated by advanced digital instruments. The Data Capacitor will provide IU, TeraGrid, national and international researchers with a unique facility for temporary, high speed, high capacity storage.
IU's Big Red supercomputer provides simulation, visualization and modeling services to accelerate the task of producing scientific breakthroughs. The S2A9550 storage cluster solution is an integral component of IU's simulation services that is capable of supporting 20 trillion floating-point operations (teraflops/TFLOPS) from just 8 S2A9550 systems. The S2A9550s are able to deliver 20GB/s high sustained I/O performance to the Big Red supercomputer, thereby reducing processing times and accelerating job completion for these scientific and research activities.
IU's Data Capacitor High-Speed Storage Repository is supported by 6 S2A9550s, delivering 14.5 GB/s of sustained throughput, easily exceeding the acceptance criteria threshold of 12.5 GB/s. Specifically, the S2A9550 is able to enlarge the data processing capability for TeraGrid and IU's scientific and research activities by delivering high sustained read and write performance enabling the Data Capacitor to make more efficient use of its storage.
With the S2A technology, the S2A9550 system delivers this performance and capacity to multiple applications in a non-proprietary solution with open solution support for a variety of cluster file systems. IU's Big Red employs IBM's GPFS file system and Data Capacitor employs CFS's Lustre file system.
"Big Red systems play a major role in the national cyberinfrastructure 'TeraGrid.' The ability of the S2A9550 to deliver optimized and very high performance I/O enables Big Red to support the goal of making U.S. scientific research the most productive in the world while enhancing our nation's ability to compete globally," said George Turner, co-manager, research and technical services, Indiana University. "Big Red systems also support a variety of IU life science applications, so the S2A9550 with its massive scalability, open system support and ability to deliver high speed I/O, helps enable a variety of Big Red supercomputing applications providing IU researchers and scientists with the opportunity to perform breakthrough science."
Big Red will play a major role in the TeraGrid, the National Science Foundation's flagship effort to create an advanced national cyberinfrastructure. Cyberinfrastructure refers to supercomputers, massive data storage systems, advanced instruments and people all connected by high speed networks, enabling new possibilities in scientific research. The National Science Foundation's goal for the TeraGrid is to make U.S. scientific research more productive and to enhance the international competitiveness of the nation's scientists. Big Red was connected to the TeraGrid on October 1st of this year and is the second fastest supercomputer connected to this innovative national grid computing system.
"Data Capacitor will play an important role in the advancement of scientific research locally and across the NSF's TeraGrid network," said Stephen Simms of IU, the project's manager. "The ability of the S2A9550 to deliver high speed concurrent read and write data access will enable TeraGrid researchers to move data between remote resources and organizations in ways that previously weren't possible. Data Capacitor will function as the center of the data life cycle from data acquisition or production to computation to visualization and eventual archive to HPSS which uses the S2A9550s for extremely fast caching. The ability of the S2A9550 to deliver high speed concurrent I/O has helped us make some significant breakthroughs in long distance data transfer," Simms said.
According to Simms, "Currently, the size of data sets that can be analyzed is limited by available computer disk capacity. IU's new Data Capacitor will allow hundreds of terabytes of storage to be dedicated to a single analysis making a considerable impact on life science researchers, where data management challenges are particularly severe. Further, Data Capacitor will allow researchers of many disciplines the improved ability to draw meaning from their data, helping them to develop new insights that result from the ability to better analyze larger data sets than can be feasibly done today."
The S2A9550 solution will be attached to both IU programs and are expected to be ranked amongst the world's most powerful supercomputer systems. Big Red will increase the University's computational capability by 20 TFLOPS making it the largest university computing centers in the U.S.
Big Red incorporates 512-node IBM e1350 BladeCenter Cluster with dual processors utilizing dual-core PowerPC 990 2.5 GHz processors, and IBM's GPFS file system all supported by 8 DataDirect S2A9550s.
Data Capacitor is comprised of 52 of Dell's 9th generation servers, each with 3.0 GHz Intel Woodcrest dual-core processors. The CFS supported Lustre file system runs on PowerEdge 2950s - 24 Object Storage Servers (OSSs), 2 Metadata servers (MDSs), with two additional servers for cluster monitoring and management. Furthermore, 24 Dell PowerEdge 1950s will function as data transfer servers.
The DataDirect S2A9550 is based on its Silicon Storage Appliance (S2A) technology, a 7th generation RAID storage networking system, which delivers up to 3 GB/s and support up to 840 TBs. The system supports Fibre Channel-4 (4 Gb/s) and InfiniBand-4X (10 Gb/s) host connectivity simultaneously, 1,120 Fibre Channel or SATA disks and works with a wide range of shared and parallel file systems. The S2A9550 can deliver 720 TB in two racks, offering the highest performance and most dense solution with the smallest footprint in the industry. Further, the S2A provides enterprise-class data protection and reliability with multiple levels of redundancy, incorporating on-the-fly parity checking of all read I/Os and Hardware RAID 6 with no loss of performance across a broad range of computational, visualization, and nearline environments. DataDirect Networks' S2A offers broad infrastructure support powering compute clusters from IBM, Dell, HP, Cray, SGI, Bull and others.
"We are very pleased to have been selected to be a part of the groundbreaking IU program that will bring new capabilities for life scientists," noted Alex Bouzari, CEO of DataDirect Networks. "Our leadership role as the largest supplier of high performance storage to more than half of the world's top 100 supercomputing sites, enabling 5 out of the top 6 fastest computers in the world, coupled with the overwhelming market acceptance of our S2A storage technology from programs like the Big Red, Data Capacitor and TeraGrid initiative, validates our decade long commitment to providing the best-in-class scalable storage solutions for high performance computing environments."