The time-worn Network File
System (NFS), designed by Sun Microsystems Inc. of Santa Clara, CA, over twenty years
ago has been getting a much needed makeover. NFS, of course, helps systems to connect so
that you can obtain a single view of all files in an environment. But as it was designed
for networks circa 1984, its performance left a lot to be desired.
Enter NFSv4.1 which includes a key component known as parallel NFS (pNFS). pNFS
enables high-speed data movement between machines.
“The most significant component of NFS V4.1 is the inclusion of pNFS which represents
the first major performance upgrade to NFS in over a decade,” said Matt Reid, director of
product marketing at Panasas Inc. of Fremont, CA.
“pNFS represents the standardization of parallel IO and allows clients to access storage
devices directly and in parallel thus eliminating the scalability and performance issues
associated with NFS servers deployed today.”
pNFS let’s you do several things. You can stripe a single file across multiple NFS
servers, which is essentially the same as RAID0. While RAID0 boosts performance by
allowing multiple disk drives to serve up data to in parallel, pNFS takes the concept and
extends it to multiple storage devices connected to the NFS client over a network. Even
if files are too small to stripe, those files can be distributed across multiple NFS
servers. As a result, it provides statistical load balancing. Further, with a capable
cluster of NFS servers and a backend file system, files or ranges within files can be
relocated transparent to the applications accessing data over pNFS.
This is achieved by the separation of data and metadata, i.e. pNFS moves the metadata
server out of the data transfer path. It is actually the same thing that Panasas does in
its ActiveStor Parallel Storage Clusters.
At its heart is the third generation ActiveScale OS with the PanFS parallel file
system which turns files into smart data objects and then dynamically distributes data
transfer operations across Panasas StorageBlade modules. This object-based storage
clustering architecture enables parallel data paths between StorageBlade modules and
client nodes. According to Reid, the ActiveScale OS dramatically lowers the cost of
managing data storage by supporting a massive expansion in data capacity within a single,
easily managed namespace.
“Panasas has been a leader in the development of the pNFS standard from the very
beginning,” said Reid.
Panasas CTO Garth Gibson, for example, co-authored the standard’s initial problem
statement in 2004. Since then, the company has provided substantial technical input to
the NFSv4 committee working on pNFS. In addition, it is directly contributing source code
to the Linux NFS client and server as well as to the Linux object storage driver, iSCSI
driver, and SCSI stack.
Most importantly, the architecture for the pNFS proposal was derived from the Panasas
DirectFLOW parallel protocol, which is a core component utilized by the Panasas PanFS
parallel file system.
“The Panasas DirectFLOW protocol used in our ActiveStor Parallel Storage Clusters
currently provides essentially all of the functionality expected to be available in the
NFSv4.1 later this year,” said Reid.
That leads him to assert that his company offers the smoothest transition to fully
compliant pNFS-compatible storage systems. Panasas and a host of other major storage
vendors will begin shipping such systems next year.
“Users and application developers can take immediate advantage of the high performance
and superior ease of use of the only production-proven parallel storage system today, the
ActiveStor Parallel Storage Clusters from Panasas,” said Reid.
Panasas, of course, has been shipping parallel storage systems to Fortune 500
companies and major government institutions since 2003. It sees the fact that its
architecture has essentially been incorporated into the NFS protocol as a significant
“We are very excited that the Panasas architecture is being adopted as standard by the
storage industry as it validates the importance of parallel storage and our vision for
the future of file storage,” said Reid. “NFS V4.1 will finally give users a reason to
aggressively migrate to the latest version of NFS, primarily because of the dramatic
performance benefits of pNFS.”
This is good news for the user community as a whole. After all, there are a number of
competing proprietary and open-source parallel file systems in existence. Standardization
of parallel IO through pNFS will allow users to reap the benefits of parallel storage
systems by choosing best-of-breed solutions without having concerns over vendor
“pNFS will accelerate the adoption of parallel storage to a broad range of customers,
from traditional HPC, to commercial HPC, to web infrastructure and media distribution,”
said Reid. “The future of file storage is parallel and NFS V 4.1 is the start of this
major industry transformation.”
Article courtesy of Enterprise IT Planet