As data storage prices continue to decline, data-rich companies must build and protect their most precious asset -- information. But like most blessings, there's a catch, and today's inexpensive storage poses some challenges that trap unsuspecting administrators. Storage is very easy to buy and add to any network, but its poor integration into an infrastructure can create storage sprawl. These data storage methods can help with data storage management.
Options for data storage management
Initially, storage sprawl may not seem that bad: If your organization needs 45 TB of storage and there's 70 TB available, who cares how it’s organized? But you’ll care when you get the bill for the management overhead and have to implement complex IT and business practices just to get to the stored data. A better option is to consolidate your data storage infrastructure, which offers far less network complexity and much less time and fewer resources spent on administration.
Because of its explosive growth, storage was once handled separately from IT infrastructure for a while. But data is now no longer held within the same servers that applications reside -- applications live on their own physical or virtual machines and their computing muscle is devoted to application efficiency. Storage resides elsewhere in the network, within its own subsystems or infrastructure. Typically, there are three main data storage management approaches: network-attached storage (NAS), storage area network (SAN) and cloud storage.
NAS devices attach to an existing network and act as giant hard drives for data. These devices store data using your preferred file system architecture. Because of their ease of deployment and administration, they are flexible for storage needs. However, because NAS appliances are so easy to deploy, it's also easy for them to proliferate on your network, and the more you have, the more challenging they become to manage.
Conversely, SANs store data as raw block data instead of using a file system. SANs can have unlimited economies of scale by tapping into a centralized storage infrastructure from wherever a user is accessing data. Setting up and managing a SAN, however, can create big expenses and management overhead.
Of the three options for managing storage sprawl, cloud storage is the newest. Not only is data provisioned here, but the virtual machines acting as file servers are also provisioned on an as-needed basis. An advantage of file servers is that they use operating systems (OSes) and file systems that match the ones currently in the existing infrastructure.
Your organization or the cloud provider can manage cloud storage, depending on your level of trust and how much you want to pay. When it comes to public or hybrid clouds, many administrators are hesitant to completely entrust their data to a third party. Using a hybrid or private cloud model, though, entails higher administrative costs.
Data storage management considerations and best practices
In all three of the above methods, the biggest setback is management overhead. Proper planning is essential to any good IT infrastructure, and storage projects are no exception. As your business grows, for example, how will you scale your storage, and which policies will prevent storage sprawl?
Storage is also dynamic over time, so you can’t expect your initial plan to remain the same. As your business needs change, so will your storage needs. Use software tools to track your storage use, analyze that data to identify trouble spots or storage bottlenecks, then implement improved procedures semi-annually.
Standardize your storage software and hardware wherever you can. Whether using NAS, SAN or cloud storage, a simplified approach to hardware and software will help keep your management costs down. Deploying multiple different storage arrays from different vendors, for example, is often more expensive and cumbersome to maintain.
Another area to reduce complexity is backup infrastructure. Moving your backup storage onto a single network location will reduce the total amount of hardware and software needed for storage and save on administration costs.
Consolidating your file and block storage assets can also bring big savings. SANs use block data storage and are highly efficient in storing and retrieving data. But users have OSes that see data as files within file systems, which makes NAS easier for many applications. If deploying a SAN, however, one barrier is translating the block data into usable files.
Historically, one fix to this problem was to install a NAS gateway appliance to the network that serves as the bridge to the SAN. NAS gateways have no storage of their own, but act solely to translate data back and forth from the file system to block data formats. Properly deployed, NAS gateways consolidate file and block storage assets in a network, and users can access files regardless of how they are stored.
A recent development is a hybridization of file and block data storage services using network unified storage (NUS) devices. NUS blends NAS and SAN protocols into the same subsystem, similar to a NAS gateway. But unlike a NAS gateway, NUS appliances have their own onboard storage. Therefore, you have storage protocols and storage space in the same device, reducing management requirements.
There are some caveats to NAS gateways -- they have their own specialized OSes, making it difficult to properly integrate them into your network. There is also an additional layer of management for maintaining these systems. Before implementing NAS gateways, weigh the advantages of deployment and be sure the payoff will be worth it.
Virtual NAS appliances
Virtual NAS appliances are new solutions for data storage management and consolidation. They give you the advantages inherent with a virtual file server, such as a familiar OS and file system, and may use a specialized NAS device. Often, the appliance just uses software (instead of a specialized device) to emulate a NAS to an available storage subsystem. Virtual NAS machines can serve more users with less memory as memory is able to expand as needed. Virtual NAS appliances can also move to physical machines as needed to provide optimal performance.
The flexibility of a virtual NAS appliance gives organizations many storage options. In a case study of its client, Nasuni Corp., Storage Switzerland detailed how Nasuni uses a fixed-size virtual NAS to hold active data as a cache, automatically moving older data out to the cloud. If a user needs older data, it’s pulled from the cloud with just a slightly longer access time.
This case study highlights how administrative overhead is reduced when storage space in the cache is fixed and does not need hands-on management. Only the archived storage in the cloud does, and since it is older data, it’s not a high priority to maintain. Virtual NAS gateways are a creative solution and make it easier to automate your storage management policies.
Whether using file server machines or virtual NAS devices, virtualization is also a technology to consider when consolidating storage. Provisioning storage appliances, too, as needed, is a good data storage management option -- as long as it is paired with well-conceived storage policies and business procedures.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at firstname.lastname@example.org.