As cloud-based solutions for backup and sharing are being offered to users, the amount of storage needed for cloud services keeps increasing. In order to lower the costs (e.g., hardware, energy) for operating such systems, it is important to rely on efficient paradigms. Currently, many systems still rely on well-proven replication to provide high availability from non-reliable devices. While easy to understand and implement, replication is far from being optimal with respect to the trade-off between storage and availability. Arguably, erasure correcting codes can significantly lower the amount of storage needed in data-centers.