NAS vs. SAN: Which tool for the job?

If you’re like me, you’ve probably tried at one time or another to drive a nail with the flat side of a wrench or attempted to remove a bolt with a hammer. If you haven’t, I’m here to tell you: Don’t do it. Listen to my grandfather, who was perhaps a little too fond of the cliche: Use the right tool for the job.

Sometimes, the right choice is not so obvious. Confusion reigns, for example, about the best applications for NAS appliances in the enterprise — and how they compare to SANs. Adding to that confusion are a growing number of devices that offer both NAS and traditional SAN capability in the same box. As you build your organization’s storage strategy, it’s important to understand the distinction between the two.

As they say: Never bring a NAS to a SAN fight (or vice versa).

Drawing the line between NAS and SAN

If you know a little about enterprise storage, you probably understand the conventional distinction between NAS and SAN: A NAS device works at the file level, while a SAN operates at the block level. File systems organize raw blocks of binary data in such a way that you can create identifiable files with human-recognizable names. Depending upon the file system in use, a NAS also lets you perform more advanced tasks, such as enforce security constraints, specify usage quotas, perform file indexing, and the like. From there, your operating system (or your NAS appliance’s operating system) allows those files to be shared on the network for clients to access.

SANs, on the other hand, make it possible to share a raw, block-level disk over a network. The composition of that network has generally made the distinction between NAS and SAN easy to spot, because most SANs have used Fibre Channel, a purpose-built storage networking protocol. But with the increasing use of iSCSI, FCIP, and FCoE — which allow block-level disk access over standard TCP/IP or Ethernet networks — this distinction has blurred significantly. Even though the actual implementation of a SAN is generally far more complex than that of a NAS, the concept is more basic because the storage is being shared at a much lower, more fundamental level.

When to choose NAS

As you might imagine, NAS appliances are great for sharing files to networked clients. That’s precisely what they’re designed for. Most of these devices target small businesses and offer features such as multiprotocol file sharing, some form of built-in backup, and such enhanced file management tools as advanced quotas and utilization reporting. If this is what you need, a NAS appliance may serve you well.

At the same time, many NAS appliances also provide some SAN-like functionality. Given that iSCSI requires very little in terms of dedicated network infrastructure or hardware — most times an industry-standard gigabit network card and a gigabit switch will do the job — there is a very low barrier to supporting iSCSI in software. That’s why many NAS devices often advertise their ability to act as an iSCSI SAN as well.

But there’s a big difference. Most iSCSI-capable NAS devices create a large file on one of their already formatted file systems and treat that file as a raw, block-level device. This introduces a few extra steps that the NAS must complete anytime it wants to read or write from that volume and can result in comparatively poor performance. Additionally, NAS appliances generally do not include the same level of hardware redundancy or high-availability functionality as a SAN, making them less suitable for large enterprise needs.

When to choose SAN

SANs provide high-performance access to block level storage on many servers at once. There are many benefits to implementing a SAN, among them increased reliability through the use of highly redundant components and consolidation-related cost savings. They are perfect for centralizing the raw disk requirements of demanding server applications such as databases, mail servers, and virtualization hosts. If you plan to do any of these things, a SAN is generally what you want.

If you want to have the best of both worlds — high-performance shared disk and dedicated, feature-rich file sharing — fear not. You can. NAS gateways are a relatively new breed of device, which fill the same niche that a general-purpose, SAN-attached file server would, but with more advanced user management and improved performance. The difference between a NAS and a NAS gateway is that the NAS gateway will generally not have any storage of its own, instead using your SAN to provide the raw storage space and focusing on the task of intelligently serving files across the network.

The only thing to watch out for with NAS gateways is that they frequently offer features that overlap with capabilities you may have already paid for as part of your SAN platform (or server virtualization platform), such as remote replication or high availability. In most cases, however, the feature overlap is preferable to maintaining two separate pools of storage, each with their own data protection and capacity concerns if NAS-related features are desired.

Before someone beats me to it, I should probably also point out that there is a very small but growing marketplace of very good enterprise-class SANs that also have full-featured enterprise-class NAS functionality without the need to implement a separate gateway. In fact, it’s so small I can think of only one off the top of my head. And unless they send me a t-shirt or something, I’ll leave it up to you to figure out which one I’m talking about.

The bottom line here is that you should not buy a NAS because it has SAN functionality. If you find yourself in the market for a SAN, buy a SAN and use a SAN-attached file server or, in situations that call for it, a dedicated NAS gateway to share your files. Make my grandad proud: Use the right tool for the job.

This story, “NAS vs. SAN: Which tool for the job?,” was originally published at InfoWorld.com. Follow the latest developments in storage at InfoWorld.com.