Is the Time Right for Network Storage?

A data backup and recovery strategy means never having to say, “Oh no! Now what?” It’s not exactly stirring up controversy to say that every business needs a plan to rescue important data should disaster strike. However, what isn’t so obvious is the best way to approach data storage itself.

Most small and mid-sized businesses (SMBs) have relied on storage devices that attach directly to PCs or file servers (sometimes referred to as Direct Attached Storage or DAS). However, network storage products that offer increased data availability in the event of failure and scalability to allow for growth — products that were once the domain of large enterprises — are now within the reach of SMBs.

As the term implies, network storage means storing data on the network — simple enough, a traditional file server covers that — but new configurations based on NAS (Network Attached Storage) and SAN (Storage Area Networks) offer more accessible and more manageable data than that offered by hard disks directly attached to a file server.

Know Your NAS From Your SAN
You might be thinking that your business is too small to need network-based storage. However, even if you have just a handful of desktop PCs and a single server, you can benefit from a NAS. Vince Gayman, director of Worldwide SMB Product Programs at HP, said a NAS is often the ideal second server, because it can offload storage requirements from the primary server and because NAS is optimized for file sharing.

Think of a NAS device like a hard disk or disk array that attaches directly to the network, without requiring a server computer. That makes a NAS simple to install and operate. Inside the box, a NAS is really a server computer optimized for managing storage, often running a trimmed-down operating system that can make a NAS faster than a general-purpose server. For users on the network, a NAS device looks like another logical drive, even though these users may be running different operating systems. Another advantage is that NAS provides centralization and a natural point of backup for shared data.

Moving up the storage food chain, a SAN is a collection of storage devices (that might include tape libraries or optical storage for backup) connected to servers through a second network, which operates separate from the primary LAN. SANs most often operate over very high-speed Fibre Channel networks (more on Fibre Channel below).

Unlike a NAS device, which stores and transfers data organized into files, SANs operate at a lower level and can transfer blocks of data across the network. That allows a SAN to move large amounts of data very quickly, offer near-infinite scaling and allow the distribution of data from multiple servers across multiple storage devices. This makes a SAN the best storage choice for high-availability, mission critical storage, but the cost and complexity of flexible storage topologies running across fiber may understandably give many SMB pause.

Why Move to Network Storage?
Different SMBs might first consider network storage for different reasons. One often-cited driver is compliance with regulations such as Sarbanes-Oxley, which mandates long-term storage of e-mail and other company data, or HIPAA, which requires stringent privacy protection. But, according to Michael Parker, product marketing manager at Veritas, compliance on its own is not enough to tip the balance. “It’s not just regulatory compliance,” says Parker, “[SMBs] are asking for more, almost regardless of size.”

Another factor is the importance of applications that might not have been central to the operation of a business a few years ago. “E-mail to many, many businesses may be a critical application,” says HP’s Gayman. And, of course, tolerance for Web-site downtime is decreasing as more and more communication and commerce funnels through the Web.

Network storage addresses these issues by making backup operations faster and more easily automated, and, in the case of SANs, by increasing data availability in the case of failure. For example, RAID (Redundant Arrays of Independent Disks) arrays directly attached to a file or application server can increase data protection and availability through disk mirroring, or through distribution of data across drives in the array. But the main advantage of a SAN, Gayman points out, is that storage is independent of the file server, allowing fail-over to a second server should the primary server go down.

Backing up NAS units on a network is much like backing up individual file servers. You can back them up one-by-one to dedicated tape devices, or you can back them up across the network using automated backup software. Storage devices on a SAN can be logically grouped and backed up to archival devices on the SAN itself.

Trends Favor Network Storage
While network storage has typically been associated with large enterprises, new technologies have brought down the cost, while storage vendors increasingly court the SMB market. Technological innovations include Serial ATA (SATA), a less-expensive alternative to SCSI for connecting individual drives (for example, within an array). SATA devices don’t have the speed of SCSI, let alone Fibre Channel, but they offer tremendous cost savings. Those savings, of course, apply whether you integrate your SATA array as a NAS, as part of a SAN or directly connected to a server. But the lower speed of SATA makes it more appealing as secondary storage as part of a networked solution rather than directly connected as primary storage on a mission-critical server.

While SATA is relatively slow but affordable, Fibre Channel makes SANs fast, but its high cost may put a SAN out of the reach of a smaller business. Offering the best of worlds is Internet SCSI (iSCSI), a lower-cost alternative to Fibre Channel that can play a part in a lower-end SAN. iSCSI allows SCSI drives and controllers to communicate using TCP/IP. That means iSCSI can be run over your existing Ethernet cabling rather than requiring you to add a specialized Fibre Channel network. Because it uses the protocol of the Internet, iSCSI devices can also communicate with remote devices across the Internet.

Even Fibre Channel SANs are growing less expensive as vendors make a move into the SMB market. HP, for example, offers a sub-$10,000 “SAN-in-a-box” (which includes host adapters, cabling, Fibre Channel switch, and management software, but no drives) and touts both its affordability and ease-of-use. “Simplification has been a major push for us,” says Gayman, discussing the company’s SMB offerings.

Josh Howard, storage specialist at CDW, agrees that falling prices and simplified bundles have made a difference. With lower prices and easier-to-use configurations, he says, “we’re able to help [SMBs] get into [storage solutions] they wouldn’t have thought about two years ago.”

The Bottom Line: Is the Time Right?
While the benefits of NAS devices are becoming clear for many SMBs, the bar for getting the most out of a SAN is somewhat higher. That said, SANs are worth considering even for smaller businesses with a need for high-availability servers — that is, if you can’t affordable to lose access to data for any length of time. “As soon as you have multiple servers that need to run in a fail-over environment, you can take advantage of a SAN,” says Gayman, adding, “multiple [servers] can be two.”

Your storage strategy needs to include an implementation plan that grows from existing direct-attached storage to network storage. CDW’s Howard suggests planning a move to network storage at the time you consider your next server purchase. Any resource limitation that you come up against, be it space on the primary server or a shrinking backup window, should prompt you to consider network storage.

Once you’ve made the move to network storage, growing with your data requirements is easier than if your storage is of the direct-attached variety. That’s because NAS devices are easy to add and SANs will scale up from starter kits to enterprise levels. Michael Parker from Veritas counsels moving forward by automating data protection as much as possible and then by centralizing data access on the network.

Should something as complex as a SAN be avoided if the business owner or office manager doubles as the IT department? No, suggests Howard, but he cautions that small businesses should get advice, because configuring SANs “is something you don’t really learn on the job.” However, he adds that what really should be avoided, if you don’t have an IT staff, is “anything that requires daily maintenance,” such as changing backup tapes or backing up clients manually — and that’s where network storage comes in. While it may be more complicated to set up initially, network storage can help to automate those tasks.

Steve Apiki is a freelance writer and software developer who works for a small business in Peterborough, NH. He’s been a contributing editor to BYTE and to FamilyPC.

Do you have a comment or question about this article or other small business topics in general? Speak out in the Forums. Join the discussion today!

Must Read

Get the Free Newsletter!

Subscribe to Daily Tech Insider for top news, trends, and analysis.