It's long been the case that large enterprises cannot operate without their digital information assets. Whether its sales records, designs, video clips, security logs, or experimental results, the amount of data has become so prodigious that dealing with it electronically is the only practical option. Electronic data is so important to the conduct of business that virtually no large enterprise lacks professional-quality electronic data storage, along with fault and disaster recovery and data protection strategies.
Despite their difference in size and scale, SMBs need much the same attributes in a storage system as their large enterprise counterparts, namely:
- High (terabyte+) capacity and performance
- Multiple client types (e.g., Windows and *nix)
- Both block (SAN) and file (NAS) access modes
- Recoverability from failures and disasters
- Low cost of ownership, including both purchase and operating cost
For SMBs, there is one additional, yet critical, requirement that often overshadows all of these: ease of use. Large organizations typically have highly-skilled staffs dedicated to managing all aspects of their IT operations, including data storage. For most SMBs, this level of specialization is not affordable.
Unified Storage Systems
Thus, storage system ease of use is typically the number one criteria for SMBs when evaluating solutions, sometimes even overshadowing performance and resiliency.
The introduction of Internet SCSI (iSCSI) has enabled the storage industry to take a giant step towards both low cost of ownership and ease of use with a generation of unified data storage systems that deliver both block and file services over TCP/IP networks. Most unified storage systems from traditional, legacy vendors are "built-up" by configuring ordinary servers with RAID adapters and NICs, and installing Microsoft's Windows Unified Data Storage Server (WUDSS) software. Others are purpose-built from the ground up in order to remove much of the complexity from the user.
Building a unified storage system by integrating existing parts is a fast path to market, but it is usually not a wise choice for SMBs without extensive in-house technical expertise. To deploy such a system, one must master disk, RAID, operating system, network, and file system configuration, along with a host of other technical details. Moreover, each component in a built-up system typically has its own user interface. Administrators are faced with different management paradigms for RAID, operating system, network, and the storage server itself. None of this is a recipe for administrative simplicity.
What is Ease of Use in an SMB Storage System?
If two data storage systems meet an SMB's needs for capacity, recoverability, performance, and cost, ease of use is the obvious differentiator. Typically, an SMB's storage system needs to be a "set and forget" device. Operations like file system and virtual device creation are relatively infrequent, and thus not likely to become second nature. Automation and "foolproofing" of these tasks are therefore particularly important.
Ease of use starts with the out-of-box experience. While initial setup is unlikely to be performed more than once, an easy installation gets a system up and productive quickly. Purpose-built systems tend to outscore built-up ones on this criterion, since there is no operating system installation, RAID adapter configuration, and so forth.
The fundamental storage system administrative operation is the creation and presentation of virtual devices (LUNs) and file systems. Most large enterprise storage systems offer a variety of RAID, mirroring, and multi-pathing options. While probably useful in large enterprises with dedicated IT staffs, the learning (and re-learning) curves for these options can make them more of a burden than a benefit for SMBs. In the absence of dedicated IT personnel, systems that default intelligently and minimize or eliminate the need for obscure technical decisions are generally preferable.
With built-up storage systems, one typically selects disks, configures and initializes RAID groups, and creates file systems or virtual LUNs, often using completely different administrative interfaces for each task. Making the virtual storage available to clients is an additional step. Purpose-built systems, on the other hand, tend to combine these tasks under a single interface, even eliminating some.
Modern storage systems are quite self-reliant. Most can compensate for disk and other hardware component failures automatically, without missing a beat. But it's important that administrators learn of failures or other conditions like over-temperature that can threaten data availability. For SMBs, most of which don't have data center management frameworks, proactive notification of failures, for example, via email, is key.
Finally, SMBs are increasingly facing needs for advanced storage system features like snapshots to protect against data corruption, and replication for disaster recovery. Again, these features tend to be easier to configure and use in purpose-built systems than in built-up systems, where administration may require manipulation of several separate components via different user interfaces.
In summary, when an SMB decides it's time to consider "industrial grade" data storage, ease of use should be one of the most important evaluation criteria. Installation, configuration and presentation, fault reporting and management, and access to advanced features should all be considered in the context of skills and capabilities available to the SMB.
What To Look For In Your First Enterprise-Class Storage System
Installation |
Number of installation steps prior to management console accessibility |
|
Configuration |
Number of parameters and decisions required to create and present a file system or virtual LUN |
|
Fault management |
Automatic fault reporting mechanisms |
|
Advanced features |
Availability of snapshots, replication, and other advanced features |
Paul Massiglia is Chief Technology Strategist, agami Systems, Inc.
www.agami.com
