FastScale: Accelerate RAID Scaling by Minimizing Data Migration

Previous approaches to RAID scaling either require a very large amount of data to be migrated, or cannot tolerate multiple disk additions without resulting in disk imbalance. In this paper, we propose a new approach to RAID-0 scaling called FastScale. First, FastScale minimizes data migration, while maintaining a uniform data distribution. With a new and elastic addressing function, it moves only enough data blocks from old disks to fill an appropriate fraction of new disks without migrating data among old disks. Second, FastScale optimizes data migration with two techniques: (1) it accesses multiple physically successive blocks via a single I/O, and (2) it records data migration lazily to minimize the number of metadata writes without compromising data consistency. Using several real system disk traces, our experiments show that compared with SLAS, one of the most efficient traditional approaches, FastScale can reduce redistribution time by up to 86.06% with smaller maximum response time of user I/Os. The experiments also illustrate that the performance of the RAID-0 scaled using FastScale is almost identical with that of the round-robin RAID-0.