Network-attached storage (NAS) has many benefits for storing and managing files, and is becoming increasingly important as unstructured data outgrows structured data across large and small organizations. But
You can read of a transcript of the interview below or download it as an MP3.
Download for later:
Get a handle on NAS sprawl
• Internet Explorer: Right Click > Save Target As
• Firefox: Right Click > Save Link As
SearchStorage.com: What is NAS sprawl, and why is it a problem?
Boles: It's really the issue of having more than one NAS unit and how you lose the efficiencies of management that you get in this NAS appliance as you start to add more of them.
The problem is that NAS storage is very attractive as a technology in the enterprise. NAS appliances have massively simplified how we manage storage for unstructured data, and even for structured data when it is applied to databases and Exchange. But as with any storage system, NAS is limited in capacity and in performance. In the face of exploding data growth, you end up adding more NAS when you exceed the capacity or performance of one particular device.
So you can easily end up with lots of [NAS devices] and that introduces a bunch of challenges. There are lots of dependencies on NAS units themselves. You might have the unstructured data or the structured data on a NAS system tied to applications that are hard to migrate, and they might be scattered across multiple NAS devices. Negotiating downtimes to do migrations and to service those NAS units might be a big problem for you. That's especially the case when you have a complex web of application dependencies between Web servers, application servers, with data scattered across a whole bunch of NAS devices.
And when those things get obsolete and you have to negotiate those downtime windows and do service events or big migrations, NAS can rapidly become a big challenge if you are working with tens or hundreds of units.
Moreover, managing NAS as a resource systematically across all these different devices can be tremendously challenging. It's hard to get a big portal into all those pools of storage that can help you manage your utilization and figure out what you have out there as far as unstructured and structured data. As you get more devices, there's more to do on a daily basis with patches and upgrades and so on.
SearchStorage.com: What can storage managers do when their NAS is out of control?
Boles: There are a couple of things you can do to address NAS sprawl. There are lots of technologies out there that you've likely heard of before under the [NAS] management category, things that can do different things that can help you get a better picture of your storage. And there are cloud services out there.
But in fact, some of the fundamentals of NAS technology are evolving and there are lots of innovators out there working on the fundamentals to address some of these challenges with systems that scale out. With scale-out NAS, you can granularly add capacity or performance so you can tune your NAS to match your needs. You can intermix multiple generations of storage nodes so that as you run into capacity or performance limitations over time, you're never faced with a migration event or buying another NAS system. You can just add more nodes to the pool of storage you have out there.
SearchStorage.com: What NAS management tools are out there?
Boles: There are some things you can overlay on top of your NAS environment -- like file virtualization, for instance -- where you can migrate files seamlessly and they look like they are still in the same location. Those can give you a unified view of your environment so you can see capacity across your environment. You can migrate files if you run into capacity limitations on a particular box, and you can see your overall data file-types because [file virtualization] unifies all these sprawling NAS devices.
But you can also get that through some information classification and management-type tools. In fact, folks like Tek-Tools [now part of SolarWinds] and others make storage resource management tools that do a lot of this as well, and I expect to find more of that functionality coming out in the U.S. market next year.
But you're basically talking about a management layer that can help you see across all of the different repositories in your entire NAS network and understand the data types that are out there. They don't necessarily resolve the sprawl problem, but they can coordinate some of those various NAS units to make them work a little bit better together and give you a holistic picture that can ease some of your [NAS] management pain.
SearchStorage.com: How can cloud storage services address NAS sprawl?
Boles: Regardless of how well we think we police our storage and make use of our storage, we end up with a lot of data that's relatively stagnant and unchanging or is largely read-only data. Many of these files can be stored on a deep cloud archive platform, like a Nirvanix storage platform or an Iron Mountain VFS [Virtual File Store] file archiving platform. So you can get some of this less-used data out of your enterprise and on to less-expensive storage.
Meanwhile, these cloud storage services can maintain localized access for a lot of this data to make it more performant. You can then reduce the amount of primary NAS storage that you're using, which is generally very expensive storage and hard to manage when it starts sprawling, and reap the benefits of this elastic repository in the cloud that can expand just about forever. And you get something that looks like a single NAS portal with unlimited capacity behind it.
So if you can make good use of that and tie it to more performant primary NAS for a much more specific set of requirements -- maybe databases and Exchange and important unstructured data -- and move all of that old stale data into the cloud [storage services] repository, you can forever expand that cloud repository and massively reduce your NAS sprawl or maybe never have NAS sprawl.
For more on NAS management:
Mastering NAS management: A network-attached storage case study
This was first published in April 2010