WOW: wise ordering for writes - combining spatial and temporal locality in non-volatile caches
FAST'05 Proceedings of the 4th conference on USENIX Conference on File and Storage Technologies - Volume 4
Matrix-Stripe-Cache-Based Contiguity Transform for Fragmented Writes in RAID-5
IEEE Transactions on Computers
AWOL: an adaptive write optimizations layer
FAST'08 Proceedings of the 6th USENIX Conference on File and Storage Technologies
CA-NFS: a congestion-aware network file system
FAST '09 Proccedings of the 7th conference on File and storage technologies
CA-NFS: A congestion-aware network file system
ACM Transactions on Storage (TOS)
| Hi-index | 0.00 |
The High-Low Water Mark destage (HLWM) algorithmis widely used to enable a cached RAID5to flush dirty datafrom its write cache to disks. It activates and deactivates adestaging process based on two time-invariant thresholdswhich are determined by cache occupancy levels. However, the opportunity exists to improve I/O throughput byadaptively changing the thresholds. This paper proposesan adaptive HLWM algorithm which dynamically changesits thresholds according to a varying I/O workload. Twothresholds are defined as the multiplication of changingrates of the cache occupancy level and the time requiredto fill and empty the cache. Performance evaluations with acached RAID5 simulator reveal that the proposed algorithmoutperforms the HLWM algorithm in terms of read responsetime, write cache hit ratio, and disk utilization.