How Nasuni Works

Download the printable version of Nasuni Technology White Paper.

Virtual Appliance Server with Unlimited Storage

Starting the Virtual Appliance
After an IT Administrator selects a cloud storage provider, or several, during setup, the Filer then grabs credentials from Nasuni so that it can communicate with the cloud directly. From this point forward, Nasuni.com is removed from the I/O path. The Nasuni Filer talks straight to the cloud providers. Data moves only between the customer’s virtual local disk and the cloud.

Virtually Unlimited Storage
When UniFS™ starts operating, both the cache and the cloud account(s) will be completely empty. As files are copied over, they are moved into the cache. After a user-defined interval, UniFS™ initiates the very first push. Since all of the data in the cache is new, the entire contents are sent to the cloud. Large files are broken into chunks. Each file or chunk is then compressed and encrypted before being sent to the cloud. Compression minimizes cloud storage growth, saving our customers money.

This is the only time the entire cache will be pushed: UniFS™ is not a backup system. We do not dump the contents of the cache with each push. Storing data in the cloud might be cheap, but even those small fees would add up if the system kept sending near duplicates of every file. Instead, after that initial push, UniFS™ only forwards the changes between the original and the most recent version. We push out only what is necessary, reducing potential storage costs.

Synchronous Snapshots and the Push
After that initial push, once the user-defined interval elapses, UniFS™ takes a new synchronous snapshot of the file tree, checking each file chunk for changes. The snapshot is an instantaneous, sub-second picture of the customer’s file system at a given moment, and it allows UniFS™ to identify new or changed data. Both new files and altered chunks of older data are tagged as dirty.

New files are chunked, and all of the dirty data is then compressed and encrypted. UniFS™ sends each encrypted chunk to the specified cloud and receives the associated keys that allow it to retrieve files in the event of a restore or a cache miss.

Once both files and directories have been pushed to the cloud, UniFS™ generates the new root directory and tears down the snapshot, ready to start all over again.