CIO - Data recovery used to be a straightforward matter of running competent data recovery software on a single disk drive. Advances in storage technology now make a number of deployment scenarios possible. Even with the best data backup practices, though, it's unlikely for a small business to have the infrastructure to keep its data perfectly synchronized.
To help small businesses be prepared should a data disaster event strike, here's a look at how the most common storage options on the market deal with data recovery.
RAID: You'll Need Software to Complement Hardware
Network Attached Storage (NAS) appliances rank among the most common storage devices that today's businesses use. They range from simple two-bay devices to 10-bay appliances that offer Storage Area Network (SAN) capabilities. Redundancy is typically implemented using Redundant Array of Independent Disks (RAID), which offers simple mirroring (RAID 1) as well as more advanced methods that strip blocks of data across multiple disks to mitigate against a single drive failure (RAID 5) or even two failed drives (RAID 6).
Using redundant storage drives doesn't make a RAID appliance immune to failure, though. In fact, the complexity introduced by striping blocks of data across more than one drive means that a catastrophic event damaging the RAID volume can make data recovery a more involved process. Specifically, don't expect to plug individual hard disk drives (HDDs) into a PC and expect to see your data there.
Fortunately, RAID recovery software can reconstruct a broken RAID array. This process entails connecting all relevant HDDs to the system and then rebuilding the data. Another less likely, but still plausible, data disaster scenario could occur when the storage controller or some other non-replaceable part within the NAS appliance fails.
While logic calls for simply purchasing the same model of the NAS and reinserting the HDDs into the new device, doing this may actually result in data loss, cautions a data recovery expert.
Even though an identical unit should theoretically work, according to Elena Pakhomova, co-founder of ReclaiMe, the difference in software version of the NAS and internal hardware modifications added over time means you could never be sure. In one scenario, she says, fully functioning HDDs were moved to an identical NAS that didn't recognize the disks for what they were. "All data was overwritten after initialization occurred."
Pakhomova suggests that businesses make a clone copy of the disks first. This ensures that the original disk set will remain accessible. Alternatively, businesses may want to connect the healthy HDDs from the NAS to a standalone computer and perform data recovery from there.
Is there anything businesses can do to preempt RAID failures? "RAID only accounts for drive failure," says Thadd Weil from the public relations team at NAS specialist Synology. "An effective backup regime ... will save users from the risk of data loss, even in a business scenario."
SSD: Buy From Those Who Understand NAND
While it will be years yet before the solid state drive eclipses the traditional HDD, SSDs are increasingly found in laptops and servers. It's therefore crucial to understand the capabilities of this once-niche storage device.
There are many SSD makers on the market today, and it's worth noting that not all SSDs are designed and built the same way. Some SSD makers buy the requisite flash memory chips from NAND manufacturers, license existing firmware and then put it all together. These manufacturers often fine-tune the firmware for better performance (and product differentiation).
According to Micron Technology, another type of SSD makers does everything in-house; this includes using existing NAND fabrication knowledge, manufacturing the flash memory chips and rolling its own firmware.
When it comes to reliability, you may be tempted to think all are equal, but Micron contends that the devil's in the details. "The endurance and reliability of NAND additionally dictates an intimate knowledge of NAND characteristics," says Doug Rollins, a senior applications engineer at Micron. "NAND varies widely depending on its density, process maturity, test knowledge, test maturity and a host of other factors."
SLC NAND chips have traditionally been considered more robust than MLC ones due to the significantly higher writes that they can endure. (Some SSD makers push this fact in your face when touting exorbitantly priced SLC-based SSDs.)
However, Rollins warns against such single-track thinking. "Don't pin your data integrity hopes on a certain type of NAND," he says, suggesting that businesses would do well to thoroughly evaluate their needs from a cost perspective.
Ultimately, vendor support may have a significant bearing on successful data recovery should an SSD drive fail. If anything, this may dissuade businesses from deploying SSDs from new or unknown brands. Don't count them out entirely, though, as they may have partners in high places.
Data Recovery Options Limited For Nonstandard Storage (For Now)
RAID arrays are the de facto standard for data storage today, but they aren't without their limitations - namely, the inability to easily add capacity without rebuilding the entire array or to utilize storage drives of different capacities. To overcome these limitations, some storage device makers have come up with their own storage implementations.
For example, Drobo makes use of an advanced system called BeyondRAID, which supports instant expansion and the use of mixed drives. In a similar vein, Synology Hybrid RAID (SHR) allows its NAS offering to utilize HDDs with different capacities in the same array. For all this flexibility and convenience, however, businesses need to know that using non-standard implementations may mean having fewer data recovery options.
Still, Drobo devices have been on the market for some time, and at least one data recovery vendor has expressed confidence in recovering smaller Drobo volumes. "SHR is effectively RAID 5 [and SHR 2 is effectively RAID 6], with additional advanced switches in a convenient GUI," Synology's Weil says. "Anyone with a background in Linux and RAID volume tools can put and SHR array back together."
Windows Server 2012's ReFS New, Novel - And Slow
Despite recent interest in cloud storage, Windows Server still plays an important role in on-premises deployments for Windows-centric networks. One innovation introduced with Windows Server 2012 is the Resilient File System (ReFS), a proprietary next-generation file system designed by Microsoft for large-storage systems.
ReFS can auto-correct data corruption once it's detected. It will also automatically isolate the damaged portion of the storage drive should its attempt to auto-correct the problem fail.
On the flip side, ReFS is still new, and the lack of documentation can impede successful data recovery should the need ever arises. Moreover, ReFS uses a "copy-on-write" method to improve its resistance to certain hardware failures.
According to ReclaiMe's Pakhomova, an inadvertent side effect is that it can be difficult to identify which version of a data to use in the event of a failure. As a result, "Recovering data from a ReFS volume takes significantly more time and CPU power than from NTFS for the same amount of data," Pakhomova says. (ReclaiMe's ReFS Recovery tool is designed to recover data from a damaged ReFS storage volume.)
Portable Storage Devices: Follow Standard Operating Procedure
A typical business is also likely to encounter flash memory cards and portable storage devices. Both are prone to failure. In general, attempting data recovery on memory cards such as SD cards or compact flash card is no different from that of a standard HDD. Simply launch your data recovery software and copy the reconstructed data files to another storage device.
The same is true for portable storage devices, though data recovery is more challenging if you use an enclosure that performs automatic encryption. Indeed, while some data recovery vendors may be able to extricate and use the encryption key from the storage chassis, this is not guaranteed.
Ultimately, the standard advice on data recovery remains the same regardless of type of media: Never write any data to them until data recovery is completed, and - in the event of a mechanical failure in an HDD-based portable storage device - disconnect it as soon as possible, before sending it to a vendor.
Paul Mah is a freelance writer and blogger who lives in Singapore. Paul has worked a number of years in various capacities within the IT industry. Paul also enjoys tinkering with tech gadgets, smartphones and networking devices. You can reach Paul at firstname.lastname@example.org and follow him on Twitter at @paulmah.
Read more about small business in CIO's Small Business Drilldown.
- Hadoop for Dummies Today, organizations in every industry are being showered with imposing quantities of new information. Along with traditional sources, many more data channels and...
- The Top Five Ways to Get Started with Big Data Despite the increased focus on big data over the past few years, most organizations are still talking about what big data is rather...
- Data Warehouse Augmentation: The Queryable Data Store While organizations have, to date, been busy exploring and experimenting, they are now beginning to focus on using big data technologies to solve...
- The IBM Big Data Platform IBM is unique in having developed an enterprise class big data platform that allows you to address the full spectrum of big data...
- Live Webcast Best Practices: How to Improve Business Continuity with Virtualization VMware solutions include a range of business continuity capabilities to help ensure availability for applications across your virtualized environment. Learn More>>
- Endpoint Data Management: Protecting the Perimeter of the Internet of Things Not surprisingly, "Internet of Things" (IoT) and Big Data present new challenges AND opportunities for enterprise IT. Teams need to harness, secure and...
- Best Practices: How to Improve Business Continuity with Virtualization VMware solutions include a range of business continuity capabilities to help ensure availability for applications across your virtualized environment. Learn More>> All Data Center White Papers | Webcasts