|
OK, fess up. You know there are two key truths about your
company's data storage: 1) Your company has faced, now faces,
or will soon face serious capacity constraints on its
application or file servers. 2) Key files are scattered across
the enterprise, often difficult to retrieve over the network,
sometimes squirreled away on individual users' desktops and
laptops. Fortunately, there's a storage technology
specifically aimed at file management headaches. That's NAS
(network attached storage). Although NAS vendors are beginning
to add block-level I/O (input/output) functionality to their
devices, NAS is primarily intended to provide file-level
access to data. Hang a NAS device off your LAN or WAN (wide
area network), and off you go, serving up files to requesting
servers, applications, and users.
But, before you go - at least, before you go too fast -
think carefully about how NAS fits into your overall
application and storage infrastructure. What file sharing
needs can it address now? And, what about the future? What
role should NAS play as your total stored data grows? Pausing
to consider a few words of implementation wisdom will prevent
you from creating your worst storage nightmare: an inflexible
infrastructure.
1. Feed The Habits Of Your Heavy Users Identify
particular applications and user groups across your enterprise
that could benefit from improved access to shared files. Ideal
candidates for NAS include workgroups that collaboratively
generate data-intensive files, such as Web content, digital
video, engineering designs, and animated simulations. NAS
devices - particularly smaller capacity snap servers - can be
relatively quick and easy to connect to the network. This
flexibility allows them to be moved around and rededicated for
temporary departmental projects. "Because NAS just plugs into
your network, you don't need to retool your architecture
whenever you want to use a NAS device to relieve the file
serving burden from a server," says Akshay Gupta, general
manager of the NAS division for Iomega (San Diego).
For broader access to shared files, Steve Rogers, VP of
marketing for Quantum Corp.'s Networked Storage Products Group
(San Jose, CA), advises large multi-facility companies to put
NAS devices in their branch locations. Having NAS on local
LANs enables remote employees to access large data-intensive
files, such as videos used for employee training. "Imagine if
every employee in every office tried to pull a piece of
streaming media from a server on the WAN. That would congest
everything," Rogers says. "Instead, you can push those large
files to NAS boxes sitting on local LANs on the edge of your
enterprise. That keeps remote offices from having to download
the file from the central data center more than once."
2. Capture Your Rogue Files You should also
identify individuals who may be creating files that are
vulnerable to being locked from access or to being lost
altogether. Employees often create and save files that exist
in single-copy form off the network. Those files need to be
put on the network so they can be accessed and backed up. NAS
can provide a safe, centralized harbor. "A lot of critical
data is on employees' laptops. In reality, that's where a lot
of work is done," says Rogers. "Any person who has been with a
company for at least a year is likely to have a GB or more of
data sitting on a laptop." Rogers advises NAS users to
supplement the hardware device with software that
automatically pushes changed files to the network during data
dumps.
Dick Vanek, president of NAS vendor Excel/Meridian Data,
Inc. (Carrollton, TX), also recommends gathering up
decentralized files. "One of the most common uses of our NAS
devices is for home directory storage for individual
employees' personal documents," he says. "Since NAS supports
multiple platforms, all employees' home directories can be put
on centralized NAS, which ensures that their files get backed
up."
3. Know When Enough Is Enough Because NAS devices
are essentially plug and play and, therefore, relatively easy
to install, companies can address immediate point-specific
storage needs. As NAS-based file serving requirements begin to
exceed the capacity of individual NAS boxes, companies don't
need to panic. They can simply plug additional NAS boxes into
the network and quickly gain more storage space. However, as
with any storage solution, companies should think carefully
about how best to manage growth. Continually adding NAS
devices to the network can be problematic. As NAS solutions
spring up in various corners of the enterprise, it becomes
difficult to support them with shared storage resources, such
as tape libraries used for centralized backup.
Companies hoping to pool the storage on multiple NAS
devices for shared use should also proceed with some caution.
"There are software packages out there that allow you to
aggregate your NAS storage, but they can be expensive," Gupta
says. "So, you have to determine what level of sophistication
you can manage. Taking NAS to the next level probably means
pooling it with other storage in your data center."
4. Don't Be In SAN Denial For customers whose
scalability requirements necessitate pooling their resources,
NAS is likely to become part of a SAN (storage area network)
solution. While SANs are designed to manage block-level I/O,
they can be used for file-level I/O with the help of a NAS
device. NAS heads (also known as NAS controllers) attached to
the SAN can serve as file sharing engines. For file-level
access, they convert block data and send it out via file
sharing protocols, such as CIFS (common Internet file system)
or NFS (network file system). "Since SANs move blocks and
blocks of data, you can't simply map a drive letter to your
SAN and pull up a Word document," Vanek explains. "But, NAS
devices can talk to the SAN, where the file data might
physically be stored. The SAN administrator can designate
blocks of data for the NAS device to use." NAS heads and
controllers are fabric aware, which allows them to be plugged
into SAN fabric switches so that SAN storage can be allocated
to them.
According to Andy Pratt, president and CEO of storage
systems integrator Unique Digital, Inc. (Houston), melding NAS
with SAN is a must for large companies, where storage
requirements tend to grow by 20% to 50% every year. Pratt
admits that stand-alone NAS works for small companies, for
departments, and for temporary special projects. But, he isn't
convinced it can efficiently handle enterprise-level growth.
"NAS' file-serving functionality makes it an important
peripheral to the block activity on a SAN. If you have extra
storage somewhere on the SAN, you can pop a NAS head onto the
SAN to make use of it," Pratt says. "If you don't have NAS
working with a SAN, you're making a mistake." |