Storage

Network-attached storage: The smart person's guide

This guide is an entry-level summary of what enterprises need to know about network-attached storage.

Image: iStock/lucadp

Network-attached storage (NAS) provides file-level access, meaning that users and applications see NAS the same as any other drive. NAS became feasible in the 1980s and commercially available in the 1990s. It can be purchased in configurations from consumer-grade (for PC backups or media storage) to enterprise clusters. This guide is an entry-level summary of enterprise options.

Executive summary

  • What it is: Traditionally, NAS is a networked box without the complexity or overhead of a file server but with far more storage capacity. It presents itself to other computers as just another node connected by Ethernet.
  • Why it matters: Traditional file servers are limited in capacity and need to be maintained like any other computer, such as with operating system patches and antivirus updates. NAS boxes generally avoid these issues; however, systems based on specialized versions of Microsoft Windows (and its potential baggage) are becoming common.
  • Who this affects: A high-end NAS system would reside in a data center, but midrange and low-end models can just as easily live in an individual department or even a remote office. Therefore, the technology is better known to non-IT workers.
  • When this is happening: NAS isn't new, but there are changes afoot in how the technology lives among other developments. It is starting to get cozy with solid-state storage, cloud options, and virtualization.
  • How to get it: NAS products are now available to the mass market. Generally, large-scale systems are sold directly by manufacturers, while mid-range versions are sold by manufacturers or resellers. Open-source software is available to build your own NAS.

What is network-attached storage?

NAS is a method of connecting large quantities of hard drives in an array and presenting them as a vast repository of file-level data as a single network node. NAS differs from storage-area networks, which are similar products that also gather large quantities of drives into one, but which typically exist on their own separate network (usually based on the fiber channel protocol) and store data at the block level.

Clustered versions of NAS can connect geographically dispersed drives into a virtual NAS. That's more akin to a SAN architecture, yet retaining file-level storage and single-point-access to other network nodes. This feature was once exotic, but is now fairly standard for high-end and midrange systems.

Additional resources:

Why does NAS matter?

As NAS systems in general get smarter, they can pick up more server-like behavior — the good, bad, and ugly. NAS' primary benefit is capacity that can rival a SAN into the petabytes, but without the complexity and expense of a fibre channel network — high-bandwidth Ethernet is all that's necessary. NAS can also be far more plug-and-play (not to say they're simple, but they are simpler than equal-capacity SANs).

Additional resources:

Who does this affect?

Large and mid-sized enterprises are very likely to already own one or more NAS systems in their data centers. However, unlike a SAN, parts of a NAS product could be managed by standard network administrators rather than storage specialists. Also, you may find NAS products residing in departments or remote offices; this sometimes leads to office managers having to become modestly familiar with the technology.

Additional resources:

When is this happening?

NAS products, as with SANs, are starting to evolve beyond just large boxes filled with hard drives. Most major NAS suppliers now offer all-flash and hybrid disk/flash products. Other modern twists on NAS include products that automatically replicate their data into cloud storage and, as discussed above, versions with advanced clustering that are essentially SANs behind a NAS head unit. Whether these are the "best of both worlds" remains to be seen.

Another trend is software-defined NAS, which allows businesses to build their own systems using commodity hardware. There's some irony to this, as business-class storage companies spent years explaining how their proprietary NAS systems were much better than homegrown file servers using independent NAS software. The low cost-per-gigabyte — or terabyte — of storage hardware is making some parties reconsider that perspective.

Additional resources:

How do I get it?

Generally, purchasing and cost-of-ownership is cheaper for a NAS than a SAN. If you're with a large organization, then contact any of the major players such as EMC (soon to be Dell), IBM, Hitachi, HP Enterprise, NetApp, or Oracle. You may be referred to a local reseller. There are many second-tier suppliers who are more willing to sell directly to smaller companies. Consider hiring consultants who can guide you through a large-scale NAS purchase, installation, and maintenance.

Remember, network-attached storage is an extremely powerful tool, so do not be fooled into complacency by the existence of consumer-grade options. NAS at the enterprise level can require much of the same storage management knowledge as its cousin the SAN.

Additional resources:

About

Evan Koblentz began covering enterprise IT during the dot-com boom times of the late 1990s. He recently published a book, "Abacus to smartphone: The evolution of mobile and portable computers".

Editor's Picks