Store more with less
Tech trends | Disk arrays offer a variety of solutions for integrators
- By Edmund X. DeJesus
- Dec 04, 2008
Agencies face similar storage challenges,
so they often ask systems
integrators how they can store
more data in a smaller space while keeping
it easily accessible.
Disk arrays are one
answer. At its most
basic, a storage disk
array is a collection of
disk drives with
cache memory, controllers,
and other supporting hardware.
The current generation of arrays also
offers a smorgasbord of speed, capacity,
data protection, security, replication and
space-saving tricks. And many products
offer savings in power and cooling ?
important considerations given the cost
of energy and increasing mandates for
At the same time, the space needed to
accommodate disk arrays is shrinking.
"You can now double, triple or quadruple
your data capacity in the same footprint
using less energy," said Greg Schulz,
founder and senior analyst at consulting
firm StorageIO Group.
Finally, many government agencies are
trying to satisfy multiple requirements. In
some cases, retrieval speed is crucial. In
other situations, particularly when data is
stored online, drive speed might be less
important because Internet connections
will determine how quickly files can be
accessed. An agency customer might
want to trade disk speed for smaller size,
lower power, better cooling or lower cost.
The agency should focus on what it is
trying to achieve, said R.B. Hooks, chief
technology officer of Sun Microsystems
Federal's Storage Group.
Fortunately, new developments in technical
protocols are helping disk arrays
provide data more easily. A trend toward
functional convergence is combining
Fibre Channel and Internet SCSI drives,
Common Internet File System and
Network File System drives,
and Fibre Attached
and Serial Advanced
drives, said Andrew
Reichman, a senior
analyst at Forrester
In other words, systems integrators
have numerous options for meeting their
customers' objectives. Do they want to
shrink their data centers? Then use high-capacity
drives. Do they need quicker
access to data? If so, recommend the
fastest products. Do they want to save
money? Avoid the biggest and speediest
models and choose the next-lowest tier of
Most disk arrays use some form of
Redundant Array of Independent Disks
technology, which achieves higher reliability
than a single disk by distributing
data across two or more drives. Also, the
technology allows the disks to send data
faster than individual drives.
Consolidating resources is a primary
goal of many agencies and departments,
partly to reduce costs and partly to simplify
"Customers often want fewer data centers
or just one," said Kyle Fitze, director
of marketing at Hewlett-Packard's
Storage Platforms Division. Given the
prevalence of concentrated storage, consolidation
has become not only feasible
but also fairly easy.REDUCTION STRATEGIES
Lean provisioning is one strategy for
doing more with less. Systems often allot
too much space to applications, which can
waste resources. Rather than reserving
storage space for an application or server,
lean provisioning makes that space reservation
conditional. For example, if two
applications each supposedly require 10T,
lean provisioning would grant each one
only 6T, with perhaps another 4T in
reserve. Those applications likely will be
content with their lean rations, and the
system will have space for other needs.
How likely is lean provisioning to help
your customer? It depends on the applications,
data and operating system.
"Typically, Windows has 25 percent to
30 percent usage, Unix has 35 percent to
45 percent usage, and mainframes about
80 percent usage," Hooks said. That leaves
a lot of wasted space.
Vendors implement lean provisioning
in various ways. For example, Sun offers
the StorageTek 9900 Software Suite for
virtualization and dynamic provisioning
and the Solaris ZFS 128-bit file system,
which doesn't consume storage until data
Data deduplication is another strategy.
It reduces or eliminates redundant information.
For example, if you send an
e-mail message with an attachment to a
co-worker, there are now two copies of
that attachment. Rather than store both
copies, deduplication tools store only one
and point to that one copy for all other
"While common in backup systems,
deduplication is now in primary systems,"
Reichman said. The technology saves on
the space and bandwidth needed to move
that data around.
In addition, data compression can
reduce the raw size of stored files.
Integrators should also consider tiered
storage, in which different tiers of disks
are used for different purposes.
"You can set the rules for which data
belongs on which tier," said Mark Peters,
an analyst at Enterprise Strategy Group.PROTECTING DATA
Most agencies are required to maintain
backups of their files, preferably at another
location. Many hard-drive arrays perform
that replication automatically.
"The idea is disaster avoidance without
disruption, not disaster recovery," Schulz
However, agencies often discover that
their disaster recovery plans don't work
when it's too late. For example, Charlotte
County, Fla., officials had to physically
move servers to a safe location during
Hurricane Charley in 2004.
"Many of our servers were running
without air conditioning, which is not
ideal," said Mark Ramsey, manager of
the county's information technology
Securing data is as important as backing
it up. Accordingly, some disk arrays
now offer the option of encrypting all data
and digitally shredding it if a drive is
removed without authorization.
"This can help any agency with issues of
compliance or chain of custody," said
Dave Egan, senior vice president of storage
at Fujitsu Computer Systems.
DriveTrust technology ? a joint effort by
Advanced Micro Devices, HP, IBM, Intel,
Microsoft, Seagate Technology and Sun
? features firmware that encrypts drives
and enables secure erasure. In addition,
manufacturers such as EMC and Fujitsu
offer encrypted drives.
However, if an agency customer chooses
drive encryption, integrators should
recommend an investment in specialized
software that keeps track of encryption
keys so the customer doesn't have to.EYE ON THE FUTURE
To accommodate future needs for
expansion, consider vendors that offer
compatible disk arrays at different levels
Fortunately, transmission protocols
are also changing. The hot one these
days is Fibre Channel over Ethernet
(FCOE). Many data centers already use
Ethernet for their TCP/IP networks and
Fibre Channel for their storage-area networks.
By adopting FCOE, agencies can
run Fibre Channel traffic over their
Ethernet connections without the need
for special Fibre Channel cabling or the
less-well-known iSCSI protocol.
"This simplifies implementation
because most staff [members] are already
familiar with Ethernet," Peters said.Edmund X. DeJesus is a freelance writer.