Research on Capacity Scaling Distribution Based on H-Code

2019 
Adding disks to RAID-6 storage system can not only increase the capacity of the storage system, but also promote the I/O parallelism of the storage system. To regain load balance between old and new data disks, RAID-6 scaling requires moving some data blocks to the newly added disks. In the existing approaches, the traditional approaches move a lot of data blocks, which results in high overhead for RAID-6 scaling. In this paper, we proposed a new scaling approach based on H-Code for RAID-6 storage system—HM6. It has the following two advantages: First, HM6 minimizes data migration. Second, when HM6 performs data migration, the data block is only moved in the same row parity chain. So, the row parity data is not updated and there is not additional row parity computational cost. Through real experimental performance analysis, HM6 reduces the number of data blocks to be moved by 75-97.9% and saves the data migration time by 86.1-93.3% comparing with traditional approaches.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    9
    References
    0
    Citations
    NaN
    KQI
    []