Skip to Main Content
We present the architecture for an disk based archival storage system and propose a new RAID scheme that is designed for "write once, read sometimes" workloads. By intertwining parity groups into a multi-dimensional RAID and improving the single disk reliability with intra-disk redundancy, the system achieves an elastic fault tolerance that can at least recover from all 3-disk failures. We do not stripe data to multiple disks and store related information to the same disks. Typically, all disks of the RAID are powered off, and for a read request, only a single disk has to be spun up, while write and rebuild processes require multiple disks. We analyzed our RAID scheme and showed that it provides a MTTDL that is orders of magnitudes higher than a set of RAID6 systems with a similar amount of disk drives. We also present and evaluate our prototype and show that it is well suited for archival scenarios.