Tech Field Day, Part One: Nimble Storage

I had the good fortune last week to spend two days in the company of a dozen independent bloggers, pundits, geeks and other thought leaders in storage, networking and virtualization at GestaltIT's Seattle tech field day. We sat, not so politely, through the usual death by PowerPoint and live demos from a series of vendors ranging from those I knew well (F5 and Compellent) to those like Nimble Storage, a start-up brave enough to come out of stealth in front of the Tech Field Day crowd.

Howard Marks

July 27, 2010

3 Min Read
Network Computing logo

I had the good fortune last week to spend two days in the company of a dozen independent bloggers, pundits, geeks and other thought leaders in storage, networking and virtualization at GestaltIT's Seattle tech field day. We sat, not so politely, through the usual death by PowerPoint and live demos from a series of vendors ranging from those I knew well (F5 and Compellent) to those like Nimble Storage, a start-up brave enough to come out of stealth in front of the Tech Field Day crowd.

While Sharon Fisher broke the Nimble Storage news as it happened here, I found their story compelling enough that a deeper dive would be useful. Nimble's primary claim to fame is the combination of flash memory as a huge read and write through cache with 1TB SATA drives to create a high performance system that can not only act as a primary iSCSI array but also store enough snapshots to replace conventional backups.

The basic CS220 combines a 640GB flash cache with 12 1TB SATA drives for 9TB of useable capacity and what Nimble's folks claim is the equivalent of 108TB of backup data. Like StorWise or even NTFS, they do LZ style compression on data before saving it. They then use their own log-based file system that always writes full stripes to the RAID back end, making much random I/O--which SATA drives don't handle well--look more like sequential I/O, which SATA drives are pretty good at.

Nimble uses redirect on write snapshots, which unlike the more common copy on write variety, should have limited impact on performance even if you retain a large number of snapshots for a long time. Of course snapshots alone aren't a replacement for backups since  an explosive power supply failure, or the like, that destroys the system also destroys the data and the snapshots. My friend W, Curtis Preston, who was also at the event, has written that a combination of snapshots and replication could replace daily backups, but it would still need a metadata index to help you find the version of a file you want to restore.

Today, Nimble doesn't quite allow you to replace your backup system. They can only replicate between two systems so you have to choose between a local copy for fast restores and a remote copy for disaster protection. They have promised cascading replication in a future version of the software.With execs on the team formerly from Data Domain, NetApp and Equallogic, Nimble has taken some of the best ideas these companies had, like the ease of use and scale out architecture from Equallogic, and redirect on write snapshots from NetApp, and combined them into a very interesting product.

A bit of a kerfuffle came out when one of the Nimble Storage folks referred to their snapshot technology as data deduplication. While the net result is similar to what you'd get with deduplicating systems like Sepaton and Exagrid that use delta differentials, there won't be any duplicate data stored from multiple backups over time. I'm not comfortable calling it dedupe. First, it doesn't eliminate duplicate data where the same data exists in multiple places. Second, it doesn't actually remove duplicates, it just doesn't save them.  As the man at the carnival says "Close, but no cigar."

Note that vendors do pay the organizers to present at Tech Field Day, and that expenses including airfare, hotel, meals and copious beverages are paid out of vendor appearance fees.  I personally received the usual amount of USB thumb drives, T-shirts and the like from the vendors as well.

About the Author(s)

Howard Marks

Network Computing Blogger

Howard Marks</strong>&nbsp;is founder and chief scientist at Deepstorage LLC, a storage consultancy and independent test lab based in Santa Fe, N.M. and concentrating on storage and data center networking. In more than 25 years of consulting, Marks has designed and implemented storage systems, networks, management systems and Internet strategies at organizations including American Express, J.P. Morgan, Borden Foods, U.S. Tobacco, BBDO Worldwide, Foxwoods Resort Casino and the State University of New York at Purchase. The testing at DeepStorage Labs is informed by that real world experience.</p><p>He has been a frequent contributor to <em>Network Computing</em>&nbsp;and&nbsp;<em>InformationWeek</em>&nbsp;since 1999 and a speaker at industry conferences including Comnet, PC Expo, Interop and Microsoft's TechEd since 1990. He is the author of&nbsp;<em>Networking Windows</em>&nbsp;and co-author of&nbsp;<em>Windows NT Unleashed</em>&nbsp;(Sams).</p><p>He is co-host, with Ray Lucchesi of the monthly Greybeards on Storage podcast where the voices of experience discuss the latest issues in the storage world with industry leaders.&nbsp; You can find the podcast at: http://www.deepstorage.net/NEW/GBoS

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox
More Insights