StorSimple Dedupes Application Data to the Cloud

I must admit that I've been seduced by the siren song of public cloud storage. The thought of infinite scalability, on demand, without huge upfront capital expense sounds great to me. Cloud storage gateways like StorSimple's address the challenges of public cloud storage usage, object APIs and latency. Like some other cloud storage gateways, StorSimple's appliances present the cloud as iSCSI block storage, but its secret sauce is in the application plug-ins.

Just mapping a common storage interface such as CIFS, NFS or iSCSI onto a back-end cloud storage service is a nifty trick. By using the cloud storage back end as a tier (as opposed to a cache) and, more importantly, by deduplicating the data before sending it across the Net to the cloud, StorSimple's approach should minimize both the connection bandwidth needed to access cloud storage and the pesky data transfer charges most cloud storage providers charge.

Using the cloud as a tier means that StorSimple's appliances are the primary storage tier for your applications, pushing data to the cloud as you take snapshots rather than constantly flushing a cache to the cloud. If you have lots of bandwidth, a caching solution should get your data off to the cloud sooner, but StorSimple's approach will ensure that the data on the cloud provider is application-consistent by using Windows Virtual Storage Service.

Of course, treating the cloud as a tier means the appliance has to make intelligent decisions about what data should be included in the locally hosted tier(s) and what data can be demoted to the cloud. 

Most storage systems that do automated tiering use block access frequency to make these decisions. Blocks that are frequently accessed are put in the fastest tier, and blocks that haven't been accessed in some time are demoted.StorSimple's application plug-ins allow their appliances to be more intelligent about data placement. The SharePoint plug-in not only puts the database on the appliance's internal flash storage and unpopular content off in the cloud, but also takes a page from Google's PageRank and keeps content that is linked to multiple SharePoint pages locally on flash or SAS disk. 

The Exchange plug-in actually makes Exchange 2010's built-in archive feature usable by determining the difference between a user's primary and archive data and storing the archive data in the cloud.

StorSimple has three appliance models, from the $15,000 flash-only 1010 (supporting up to 10TB of data) to the $50,000 7010 (which can support up to 200TB of data and fault-tolerant pairs of appliances). In addition to deduping the appliances, perform AES-256 encryption on your data to keep it safe from prying eyes.