IceBreaker, Line of Storage Solutions From Penguin Computing
For HPC
This is a Press Release edited by StorageNewsletter.com on November 24, 2010 at 3:45 pmPenguin Computing Inc. announced IceBreaker, a new line of storage solutions designed to make high-performance computing faster.
Penguin Computing’s IceBreaker is a range of scalable storage solutions
from IceBreaker NX, a no-frills direct-attached storage device, to
IceBreaker EX, a parallel file system solution designed for
high-availability, scalability and ease of management. Building on the
foundation of Penguin expertise in Linux, HPC and high speed
interconnect fabrics and applications, IceBreaker completes the Penguin
HPC solution portfolio and delivers multi-terabyte datasets to compute
nodes and applications.
"Penguin’s IceBreaker is coming at a good time," said Charles Wuischpard, Penguin Computing CEO. "Disk
I/O simply has not kept up with multicore processors and
parallelization. With this new line, we have addressed this reality with
storage solutions that greatly enhance the performance and productivity
of HPC clusters."
Penguin’s Icebreaker storage solutions include a range of products to
support a customer’s workflow requirements. In addition to multiple
capacity points from 12TB to 160TB, the IceBreaker FX and IceBreaker SX have specialized RAID-6
technology that enable the disk array performance to match the 40Gbps
throughput of an InfiniBand Quad Data Rate (QDR) fabric. By adding
Lustre distributed file system technology for large shared
infrastructure, the IceBreaker SX can support an unlimited number of
clients with fast storage.
A large private university in the Midwest installed IceBreaker FX with
their Penguin Linux cluster to improve the I/O performance of their
research application. The application generates large amounts of data
that must be offloaded to storage during the compute process and
integrated for final analysis. Using IceBreaker FX significantly
improved the performance of their research application and researchers
found they were finishing compute runs in a fraction of the time.