June 24, 2013
LEIPZIG, Germany, June 24 -- Scalable Informatics, Inc. (Scalable), a leading provider of high performance, tightly-coupled storage and compute solutions for a wide range of data, I/O, and process-intensive applications, announced the industry's fastest single storage rack I/O solution. The siCloud single rack storage system delivers record-breaking I/O performance using a tightly integrated system design.
Based on repeatable measurements, Scalable has achieved 90 GB/s and 11M IOPs performance running on a single siCloud rack with 30 siFlash server nodes and a 40GbE network infrastructure. The performance represents aggregate raw device I/O measurements for all siFlash arrays in the siCloud rack. The theoretical maximum bandwidth for this system is 96 GB/s and 11M IOPs, and continues Scalable’s track record of providing aggressively-priced high performance I/O solutions.
“The siCloud RAIDs were built as RAID5 systems, with a 32k chunk size upon a set of 12 SSDs per node,” explains Dr. Joseph Landman, CEO of Scalable Informatics. “An xfs file system was layered upon the RAID5, and aggregate I/O performance for distributed workloads was measured using the fio filesytem benchmark.”
Scalable also reports results for file system export via NFS, as well as Fraunhofer Parallel File system across 30 siFlash serve nodes in the siCloud rack. Each siFlash server performed at 329k IOPs at 2.7 GB/s per machine to its local file system, producing a usable aggregate of 9.9M IOPs from the file systems, at a sustained rate of 81 GB/s.
Scalable has also introduced the I/O Domination Discount Program, inviting customers to describe how their application will benefit from Scalable Informatics’ record-setting I/O performance. Customers needing improved I/O performance are invited to explain how using the siCoud 90 GB/s and 11M IOPs performance can relieve their I/O bottlenecks in order to receive a purchase discount.
Landman further explained, “We’re interested in applying this game-changing performance to challenging I/O problems customers are trying to solve. This level of performance opens up new ways of thinking about application performance and system design. If you're interested in applying to the I/O Domination Discount Program please send a brief proposal with full contact information to firstname.lastname@example.org.”
The I/O Domination Discount Program requires customers to provide a one-paragraph summary of how their application will benefit from the siCloud performance levels to receive up to a 5% purchase discount. Those customers who include supporting data and allow Scalable Informatics to cite I/O improvement and use their organization’s name on Scalable’s website can potentially increase the discount up to 10%. More details can found on the I/O Domination Discount Program page: https://scalableinformatics.com/domination.
Systems are available immediately. Scalable will be demonstrating their solutions at Booth 551 at the International Supercomputing Conference in Leipzig, Germany, June 17–19.
About Scalable Informatics
Scalable Informatics, Inc. is a leading provider of high performance, tightly-coupled storage and compute solutions for a wide range of data, I/O, and process-intensive applications. Their innovative design — tightly coupling storage and computing in the same unit — eliminates bottlenecks to deliver the highest performance at aggressive price points. As Scalable’s products work with all storage networking protocols and technologies, they can easily be adapted to any type of infrastructure. Their products are simple to install and support, and provide exceptional flexibility for scale-out performance. Scalable also provides consulting and development services, as well as on-demand computing and storage solutions. To learn more, visit http://scalableinformatics.com.
Source: Scalable Informatics
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?