Both VoltDB distributions - Community Edition and Enterprise Edition - maintain longevity through database snapshots that can be performed ad hoc, at user-defined frequencies or continuously. Pictures are recorded in persistent storage. Database recovery (longevity) is achieved by restoring from snapshots.
VoltDB Enterprise Edition also includes Command Logging, which ensures the longevity of transactions that occur at (usually short) snapshot intervals. The Logging command can be configured to run synchronously (100% guarantee of durability) or asynchronously (less impact on transaction delays due to the loss of some transactions during a failure). If you use asynchronous logging, you can configure the fsync window to balance latency and longevity.
In this way, both the VoltDB support longevity distributions with snapshots and Enterprise Edition provide additional longevity through Command Logging.
It should be noted that both VoltDB distributions also include built-in high availability through a multi-master synchronous function called k-security. You can support as many “wizards” of each node database as possible, and VoltDB will transparently (and synchronously) apply transactions to all relevant nodes. If one node fails, its peers simply continue to accept / process work. This tandem-style fault tolerance significantly reduces the likelihood of a failure requiring a database recovery.
More about VoltDB snapshots: http://community.voltdb.com/docs/UsingVoltDB/SaveSnapshotAuto
More information about registering VoltDB teams: http://community.voltdb.com/docs/UsingVoltDB/ChapCmdLog