Ocarina-EMC vs. NetApp: No Contest

Great news–Ocarina has announced that its solution for EMC Celerra is available immediately, offering its advanced, content-aware dedupe and compression to EMC NAS customers. While Ocarina already had a solution for EMC Celerra, this announcement means that has been admitted to the EMC Velocity Technology and ISV Program. A major step forward.

EMC has other dedupe options available for Celerra, but Ocarina gives them a distinct competitive edge against NetApp and NetApp dedupe. This EMC-specific release is also one of the more elegant implementations of Ocarina.  We use EMC’s mature FileMover interface to be able to insert Ocarina completely transparently on the Celerra. Users access their files on the Celerra through all supported protocols – including both CIFS and NFS. Like our BlueArc and HDS releases, we are called from within the file system, rather than intercepting calls on their way in to it.

Ocarina optimizes files out of band, and is only called on reads and writes when an optimized file is accessed. This means we are not in the path at all for accesses to hot files. EMC has a rigorous set of tests you have to pass to work with FileMover, and getting through those tests was a good validation exercise for Ocarina. I think EMC customers can feel very comfortable about how solid this solution is.

One technical feature worth noting is that using Ocarina to optimize a Celerra volume means that it is possible to greatly increase the logical size of that volume. Like NetApp, the current release of EMC Celerra has a 16TB volume size limit. FileMover – as the name implies – lets you move the contents of a file somewhere other than the original volume. The FileMover stub that is left behind makes it appear to applications and users that the file is still in the original volume. The way Ocarina works, we read the file from the original volume, optimize it, and store it in another volume on the same Celerra.  We’re not really moving it off the filer at all, but we are using FileMover to allow you to spread files out across multiple volumes on your Celerra.

A FileMover stub is left behind in the original volume. The stub does not take much physical space. So if you had a 16TB volume called volume A, and Ocarina started optimizing files and storing them in volumes B, C, and D, you could keep creating new files in volume A using the free space we just created. As you create those new files in volume A, we could keep optimizing and moving them. We can also simply move them, without optimizing them. This is completely policy-based.

The net effect, over time, is that a user could mount a single share, Volume A, and have direct access to much much more storage than 16TB. Let’s say we get 75% optimization on average across a set of files. That means you could store 64TB in one volume. With FileMover and the example above where we are using 3 volumes as targets, you could store 192TB in Volume A (though the file contents would actually be distributed across Volumes B, C, and D). This works extremely well for all typical NAS file data. The Celerra also supports unified storage, where Celerra volumes are used for iSCSI and for databases. And while Ocarina is not targeting our solution for those use cases yet, please do stay tuned.

And here’s a quick note that may interest those of you who are already Ocarina EMC shops–or who have a customer or client that is. If your company or client has benefited from the groundbreaking Ocarina solution, we are initiating a special program that may be of great interest to you. For more information about this new opportunity for a data reduction package at an exceptional price, contact: info@ocarinanetworks.com.