No One Screws Up NAS - Right?

Companies looking for a quick and easy way to add capacity and flexibility to their file storage resources should consider NAS (network attached storage) devices. NAS hardware plugs into the existing LAN or WAN (wide area network), immediately giving network users online disk space for storing and retrieving files. Because it resides on the network, where it can be targeted and accessed by multiple applications and user groups, NAS-based storage reduces many of the allocation limitations of direct attached server storage. According to Bakul Joshi, VP of marketing and sales for NAS vendor Ateonix Networks, Inc. (Fremont, CA), "A lot of companies reach the point where their application service is degraded because a server on a particular OS [operating system] is limited to using storage on the same platform. By contrast, NAS devices are OS-independent, so it doesn't matter what platforms the applications accessing NAS are running on. They can all access the same NAS device." Akshay Gupta, product general manager for NAS vendor Iomega (San Diego) agrees, noting, "NAS is a fast way to add storage because it's heterogeneous in terms of the various types of networks you may be using, and it avoids cross-platform file system issues."

Another key improvement over direct attached storage is NAS' ability to enable file sharing across a global enterprise. Says Jeff Hornung, VP, marketing and business development for high-end NAS provider Spinnaker Networks, Inc. (Pittsburgh), "With NAS devices, you can globally access a single file name space. If you need to scale your NAS, you can cluster devices and give them a single file space accessible from any location." Not only can NAS provide file sharing services across the enterprise, but it can also allow simultaneous access by multiple users. "People use NAS devices to share documents during collaborative projects," says Jeff Houle, sales manager at document management and NAS reseller Z-Source, Inc. (Stoughton, MA).

As Joshi recognizes, simplicity of installation is a key selling point for NAS. "Customers want to add storage to meet their growing requirements, but they don't want to disrupt their operations while they do that," he says. Since most NAS vendors design what are essentially plug and play devices, a NAS installation should not require a lot of heavy lifting by in-house IT staff. When it comes to choosing and bringing up a NAS device, however, simple doesn't necessarily mean "no brainer." There are still several key implementation issues to consider:

1. Alter Your DAS Mind-Set
As you migrate direct attached storage (DAS) to your NAS device(s), you may have to change your traditional approaches to how you think about storage. It's no longer a matter of "this box provides storage for this server." Storage resources are now shared. "NAS users need to stop thinking about storage solely in terms of the physical devices housing the storage," says Hornung. "Instead, they should start thinking about how to carve up pools of storage in logical, not physical, ways."

2. Check Your Wiring Infrastructure
With NAS devices hanging off a LAN or WAN, network performance has a direct impact on NAS performance. "Some companies are still running on 10BaseT Ethernet networks," says Houle. "While few companies run NAS on Fibre Channel and yours probably won't need to either, you'll want to at least upgrade to 100BaseT."

3. Discard Old Hard Disks
To cut costs, companies may be tempted to make current storage resources part of the NAS solution. That involves putting a NAS head, or controller, in front of existing disk arrays. As Hornung cautions, however, those disks may not be able to perform as well as shared resources. "Those older systems could have disk drive technology that is two or three or more years old," he notes. "You may still have some 4 GB drives, whereas today, on the Fibre Channel side, we're looking at 73 GB or 146 GB drives. If you plan to scale your networked storage, you'll want to move to a newer platform."

4. Don't Ditch Your Optical Discs
Many companies migrate their files to NAS devices to provide faster online access to data previously stored on robotic optical jukeboxes. Nevertheless, those jukeboxes can come in handy. "Leave the optical system there," advises Gupta. "You can still use it for archiving."

5. Push Higher Utilization
In direct attached storage environments, systems administrators are often hesitant to give users high quotas for storage utilization. Since DAS storage can't be moved around, they fear overutilization and, consequently, systems crashes. "A rule of thumb in the direct attached world is to run at no more than 50% to 60% utilization," Hornung says. "With NAS, you can run at 80% or 90% because you can re-allocate file space without disrupting applications' access to storage."

6. Monitor Usage
Despite the ability to run NAS devices at higher utilization levels than you would with DAS-based storage, don't let users run wild. Make use of quota management, file filtering, and usage reporting tools provided by the NAS vendor or available in a storage management software package. As Gupta explains, "You can filter out unwanted file types, such as large file formats and MP3 files. You can also get reports on the frequency of particular file usage. Inactive files can be migrated off to archival storage." Adds Houle, "A manager for an engineering project, for instance, can use the activity logs to see if users involved in the project are accessing necessary files in a timely fashion."

7. Fit NAS Into Backup Plans
Not all corporate data resides on the network. Much is saved on individual users' desktops and workstations. Make sure it gets backed up. "You can back up PC hard drives by using NAS devices as 'scratch servers,'" says Gupta. "Target a user group to a NAS box, give those users a set of volumes on the box, and schedule automatic backup - over the lunch hour, for example." And, no matter what you do, make sure that NAS, used for any purpose, is not the last line of defense for backup. "Back up to NAS; archive to tape - that's our message," Gupta says. "Archive to something so that all of your data isn't residing only on a hot device on the network."

8. Don't Ignore Threats To Uptime
Despite the relative simplicity of NAS design and functionality, NAS isn't invulnerable to downtime. According to Houle, "If you use a NAS device long enough, there will be a disk failure. If your service level agreement doesn't guarantee a quick enough turnaround for a disk drive replacement, at least buy a cold spare, if not a higher end warm spare."

9. Look Before You Make Your SAN Leap
Companies now comfortable with NAS may want to tackle a full-blown SAN (storage area network). They may need to supplement NAS' file-level capabilities with SAN's block-level functionality. But, SAN/NAS integration is far more complex than a NAS installation. "Companies make the mistake of deploying a SAN without first determining if they have the staffing to manage it," Joshi says. "NAS enables you to add storage quickly - just plug it into an existing Ethernet network. But, SAN requires setting up an additional network, so make sure you have the resources to handle that degree of expansion."