We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Nscale: an efficient RAID-6 online scaling via optimizing data migration.
- Authors
Xie, Ping; Yuan, Zhu; Hu, Yu
- Abstract
The emergence of novel storage medium relieves pressure caused by massive data on large-scale data centers. However, the storage cost is always a challenge that we cannot be ignored. As a trade-off between capacity and cost, RAID offers big capacity, low cost, high reliability, and flexible scaling, which occupies a large share of the storage market. Today, RAID scaling is the most frequent operation in storage systems. Nevertheless, it still has to face long scaling time and bad user experience. Therefore, we put forward an approach-Nscale for N-Code-based RAID-6 scaling. Nscale shortens the total scaling time by optimizing the data migration process and reducing the amount of data migration. Meanwhile, it ensures that the data only moves in the horizontal direction and in the same parity chain. In addition, it guarantees that the diagonal parity chain is not destroyed as much as possible. Derived from the experimental results, Nscale reduces the data migration by 81.05–92.35% and shortens the total scaling time by 54.5–62.4% under off-line. During and after the scaling process, Nscale also demonstrates excellent user average response time under different workloads, providing favorable user experience.
- Subjects
USER experience; MARKET share; ELECTRONIC data processing
- Publication
Journal of Supercomputing, 2023, Vol 79, Issue 3, p2383
- ISSN
0920-8542
- Publication type
Article
- DOI
10.1007/s11227-022-04752-5