Close this search box.

Spinning Down Production Disk Drives and 200:1 Dedupe Ratios; New Vendors Claim to Fame at NYC Storage Decisions

My visit to this fall’s Storage Decisions conference in New York City on Wednesday, September 24, was an abbreviated stay. I only had the afternoon to spend at the conference before leaving in the evening for another set of meetings the next day. So while my time was short, I did catch a couple of briefings as well as a little industry chatter.

Some of the talk on the exhibit hall floor had to do with the current crisis facing the banking industry and what that may mean for technology as a whole. One of the sentiments expressed which I generally agree with is that the financial crisis is probably not good news for the larger storage vendors at the show but likely bodes well for emerging storage technologies in the market as it will force some companies to look beyond traditional solutions.

I happened to run into Beth White who used to work at Vixel before it was purchased by Emulex a few years ago. She said that the last major downturn in the economy in 2001 – 2002 actually turned into a boon for Vixel. Prior to 2001, no one wanted to hear Vixel’s story about how its FC switch on a chip technology was more efficient than other FC solutions. But when the economy turned south, suddenly everyone was interested in more efficient and economical ways of connecting storage. Sales for Vixel’s products took off and Emulex eventually purchased them so for the following two emerging companies with whom I took briefings, their timing might be perfect.

The first of my two briefings that afternoon was with greenBytes that is bringing a new NAS/iSCSI storage system to market. I met with greenBytes’ VP of Sales and Marketing, Mat Aitkenhead who provided me a brief history of the company and then launched into an extensive dialog about its uses of openSolaris and the ZFS file system that can do some very innovative storage management functions. While I could describe these myself, I suggest you check out Stephen Foskett’s recent blog entries on greenBytes Cypress as well as the power of the ZFS file system and how Cypress uses ZFS to optimally place data on disk drives and spin down disk drives not in use.

More interesting to me was greenBytes’ claim that it can do online deduplication of production data which struck me as bit over-the-top. However Aitkenhead claimed that the secret sauce behind greenBytes’ Cypress is an algorithm that allows it to identify a unique block of data in as quickly as 3 search operations and that scale into the petabytes of data. My concern is that even if it works as greenByte claims, it has to find some enterprise that is willing to bet its business on it and then live with the consequences if it cannot deduplicate or scale as promised. Good luck with that.

Following that meeting, I had a chance to catch up with Ocarina Networks’ VP of Product Management, Carter George. Since I was already familiar with Ocarina Networks Online Storage Optimizer solution (it compresses and deduplicates like images in archived photos and movies), George just provided me with a brief update on one specific new feature that had to do with how it was eliminating the need to resave copies of photos that are nearly identical.

For instance, if someone takes your photo and you have red-eye, if you use Photoshop to edit the red out of the eye and then resave it, a second photo of like size is created even though just a few pixels were changed. To eliminate the duplicated pixels, the software decompresses images all the way to the pixel level, identifies what pixels have changed from the original to the edited copy, and then saves only the changed pixels. Obviously this saved amount of storage space will not matter for most companies but matters most for online photo sites like Flickr or Shutterfly where these space savings can add up to PBs of savings in storage.

One movie house is using this feature in the production of an upcoming 3-D movie. It estimated that to keep all of the data online in the making of the film would consume 200 PBs of storage so it only was only going to keep 30 days worth of data online. The rest it had to move off to tape and then recall it if it needed the data. Using Ocarina and only saving the changed pixels, it can keep it all online and expects to see 200:1 reduction in data during production. While Ocarina has only been in production for a couple of months, George reports that so far it seems to be working fine.


Click Here to Signup for the DCIG Newsletter!


DCIG Newsletter Signup

Thank you for your interest in DCIG research and analysis.

Please sign up for the free DCIG Newsletter to have new analysis delivered to your inbox each week.