Current data growth rates suggest we will someday hit a limit beyond which we will no longer be able to store data reliably. A large part of this problem is infoglut. How we use disk drives is also partially to blame.
[ MarioApicella's column is now a blog! Get the latest storage news from the Storage Adviser blog. ]
Too often we write data to disk as if we were storing sheets of paper, wrapping logically related information into bundles and then moving on to the next sheet and the next bundle when starting a new topic. And then, to prevent data loss, we make copies of every piece of information, often following with a second replica on removable media.
But what if we could use disk drives differently to create reliable, secure single instances of our data? Enter dispersed storage, a new approach to using disk drives as proposed by startups Cleversafe and RevStor.
As with a shredded paper document, no single data fragment can give away the whole. Because of this, dispersed storage is inherently more secure than traditional methods. Only the owner can bring the data confetti back together.
To that model, Cleversafe adds redundancy, allowing the system to recover the original data even if multiple fragments have been lost. The result is a resilient single-copy system that makes traditional competitors -- often shot down by just two broken drives -- seem frail by comparison. Moreover, Cleversafe's system is location-independent, allowing its nodes to be spread just about anywhere there is a good broadband connection.
Last week Cleversafe made new waves by releasing to the open source community the latest version of its software, which now offers access from iSCSI clients, among other significant developments. Sound intriguing? Check out Cleversafe's comprehensive glossary, download its software for a test ride, and then volunteer for the open source project, if so inclined.
As for RevStor, its flagship SANware product presents a decidedly different take on the concept of dispersed storage. SANware is targeted to address your disk-to-disk backup needs in the datacenter. By installing it on at least 10 machines (the more, the better), you can create a distributed pool of storage for replicating your important data.
You won’t need to add more storage, because SANware uses whatever disk space is available on each machine. You can, however, set up rules of behavior for each node, enabling you to limit the amount of storage and network bandwidth available to the pool.