HP StorageWorks Scalable File Share
Encyclopedia
HP StorageWorks Scalable File Share (HP SFS) is a production tested distributed parallel scalable filesystem designed to solve the I/O bandwidth challenge on large Linux clusters used in high-performance computing
applications. It is based on the open source Lustre
filesystem from Sun Microsystems
. It is composed of software, hardware storage and service from HP.
HP SFS can be used over ethernet networks but it is more typical to use a high performance interconnect like Infiniband
or Quadrics
elan4, with many hundreds of compute clients connecting to each server cluster.
On the server side, HP SFS uses clustered pairs of HP Proliant servers (typically DL380) connected to either MSA20 SATA or EVA FibreChannel RAID storage arrays. Each pair member is the primary server of a set of Lustre services (MDS or OST) and backup server for the services provided by its partner. This configuration means that full filesystem service is guaranteed as long as at least one member of each server pair is up. If one server goes down its partner will automatically take over serving the OST or MDS services that were lost. Clients won't notice that anything has happened, except that they will experience a slight delay while a connection to the new server is established (Lustre recovery).
On the client side, HP provides a set of source rpms that can be used to build the kernel modules and supporting files required to connect clients running various Linux distros to an SFS server. They also provide precompiled binary rpms for the HP XC High Performance Computing Linux compute cluster distro.
The current version of HP SFS is V2.3 which is based on Lustre 1.4.11
A new version of SFS, based on Lustre 1.6 and codenamed SFS G3, was released in late 2008. This ships on different hardware configurations than the original SFS versions described above. A variant of G3 for the original SFS hardware platforms is due to ship in early 2009.
High-performance computing
High-performance computing uses supercomputers and computer clusters to solve advanced computation problems. Today, computer systems approaching the teraflops-region are counted as HPC-computers.-Overview:...
applications. It is based on the open source Lustre
Lustre (file system)
Lustre is a massively parallel distributed file system, generally used for large scale cluster computing. The name Lustre is a portmanteau word derived from Linux and cluster...
filesystem from Sun Microsystems
Sun Microsystems
Sun Microsystems, Inc. was a company that sold :computers, computer components, :computer software, and :information technology services. Sun was founded on February 24, 1982...
. It is composed of software, hardware storage and service from HP.
HP SFS can be used over ethernet networks but it is more typical to use a high performance interconnect like Infiniband
InfiniBand
InfiniBand is a switched fabric communications link used in high-performance computing and enterprise data centers. Its features include high throughput, low latency, quality of service and failover, and it is designed to be scalable...
or Quadrics
Quadrics
Quadrics was a supercomputer company formed in 1996 as a joint venture between Alenia Spazio and the technical team from Meiko Scientific. They produced hardware and software for clustering commodity computer systems into massively parallel systems. Their highpoint was in June 2003 when six out of...
elan4, with many hundreds of compute clients connecting to each server cluster.
On the server side, HP SFS uses clustered pairs of HP Proliant servers (typically DL380) connected to either MSA20 SATA or EVA FibreChannel RAID storage arrays. Each pair member is the primary server of a set of Lustre services (MDS or OST) and backup server for the services provided by its partner. This configuration means that full filesystem service is guaranteed as long as at least one member of each server pair is up. If one server goes down its partner will automatically take over serving the OST or MDS services that were lost. Clients won't notice that anything has happened, except that they will experience a slight delay while a connection to the new server is established (Lustre recovery).
On the client side, HP provides a set of source rpms that can be used to build the kernel modules and supporting files required to connect clients running various Linux distros to an SFS server. They also provide precompiled binary rpms for the HP XC High Performance Computing Linux compute cluster distro.
The current version of HP SFS is V2.3 which is based on Lustre 1.4.11
A new version of SFS, based on Lustre 1.6 and codenamed SFS G3, was released in late 2008. This ships on different hardware configurations than the original SFS versions described above. A variant of G3 for the original SFS hardware platforms is due to ship in early 2009.