Dispersed

Dispersed volumes take the best of both distributed and replicated volumes by stripping the data across all available bricks and, at the same time, allowing redundancy. Bricks should be of the same size, as the volume suspends all writes once the smallest brick becomes full. For example, imagine a dispersed volume such as a RAID 5 or 6, where data is stripped and parity is created, allowing data to be reconstructed from the parity. While the analogy helps us to understand this type of volume, the actual process is entirely different as it uses erasure codes where data is broken into fragments. Dispersed volumes provide the right balance of performance, space, and high availability.