With tight budgets in the enterprise data storage world, the arguments for data reduction have never been more...
compelling. But what's the best way to do that if you are an SMB with limited resources?
Valdis Filks, a Gartner Inc. analyst based in Sweden, said the answer is painfully obvious: Make sure you are only backing up the things you want and not the things you don't want. In particular, he noted that many organizations have become unwitting repositories for employee videos, favorite music, and even family photos. "You probably need to do a spring cleaning to make sure you exclude that kind of data," he said. Aside from implementing policies on primary storage to reduce the accumulation of unwanted data, Filks said you can also simply omit whole categories of files when you do backups -- JPEGs for example -- preserving those that you really need on an exception basis.
To further enhance this process of elimination, Filks suggested acquiring tools from companies such as Northern Parklife Inc. and NTP Software, which can help ferret out categories of less-critical data. According to Filks, NTP and Northern do this specifically for the Microsoft Windows and network-attached storage (NAS) filer market. In addition, most storage resource management (SRM) products also provide this capability "but most of those are not applicable for the SMB space," he said.
Data deduplication for SMBs
Omitting certain data that your company doesn't need to store is the low-cost approach to saving space. Then, there's the increasingly affordable data deduplication option. "That's still not a truly low-cost solution," said Filks. However, he predicted that will begin to change within 12 months as increasingly cost-effective options come to market.
"I think it will start to become ubiquitous and dedupe will no longer be just for backup, it will be for primary storage, too." For the time being, Filks recommended looking at dedupe products that are either targeted at smaller organizations "or have that as their heritage." Examples include FalconStor Software Inc.'s Network Storage Server (NSS) and the soon-to-be-released version of Sun Microsystems Inc.'s ZFS file system.
The bottom line, he noted, is that dedupe capability is becoming standard in storage arrays and will increasingly be available even at the low-end of the market.
Offering a somewhat different take on dedupe, John Webster, an analyst at Illuminata Inc., said the first line of attack for an SMB should be its backup application. "Many already support deduplication and this will be getting more and more common," he said. The other option to consider, said Webster, is offsite backup and deduplication.
While acknowledging that SMBs may face limited choices in efforts to trim their data footprint, Arun Taneja at the Taneja Group said their functionality requirements are no different than large companies. "The pain is no different -- data loss is just as serious of an issue," he said. On a practical level, "a large company might care a great deal whether they achieved a 25:1 ratio or only a 20:1 ratio in deduplication because that delta represents a substantial amount of money whereas a similar difference at a very small company might not be worth chasing."
"If you are an SMB, don't get too hung up on chasing the very largest data reduction rates -- anything around 20:1 or better is fine," said Taneja. He also pointed out that unlike large enterprises, where there is perceived value in having a single image for all data, "single node solutions may be acceptable for smaller organizations."
In general, Taneja recommends that SMBs consider selecting data deduplication devices that are SATA based because that's a lower cost option. "Look for systems that are designed for smaller companies," he said. Exagrid Systems is an example of one such company. Data Domain, though now owned by EMC Corp., also started with a smaller business focus. "Those solutions are designed for companies that don't have a lot of IT expertise in house," he said.
Also, be sure to consider the advantages of buying a solution through a VAR rather than direct from the manufacturer. "They can be your strategic partner, located in your region and able to provide support and expertise when you need it," said Taneja.
Finally, according to Filks, data compression schemes, which encode or re-encode information using fewer bits, might seem to be an alternative or complementary approach to storage reduction via deduplication, which reduces aggregate storage by eliminating duplicate blocks or files. However, although some vendors have advertised very high compression rates, Filks said the process won't be much help to SMBs. "Compression really only helps if your data is suitable. Word files, for instance, can be readily compressed but Excel and PowerPoint documents often do not compress well," he said.
About this author: Alan Earls is a Boston-area freelance writer focused on business and technology, particularly data storage.