Competitive randomized algorithms for non-uniform problems
SODA '90 Proceedings of the first annual ACM-SIAM symposium on Discrete algorithms
File Assignment in Parallel I/O Systems with Minimal Variance of Service Time
IEEE Transactions on Computers
Latency effects of system level power management algorithms
Proceedings of the 2000 IEEE/ACM international conference on Computer-aided design
Massive arrays of idle disks for storage archives
Proceedings of the 2002 ACM/IEEE conference on Supercomputing
Characteristics of WWW Client-based Traces
Characteristics of WWW Client-based Traces
Proceedings of the conference on Design, automation and test in Europe
Energy conservation techniques for disk array-based servers
Proceedings of the 18th annual international conference on Supercomputing
Modeling Hard-Disk Power Consumption
FAST '03 Proceedings of the 2nd USENIX Conference on File and Storage Technologies
Queue - Multiprocessors
Hibernator: helping disk arrays sleep through the winter
Proceedings of the twentieth ACM symposium on Operating systems principles
JouleSort: a balanced energy-efficiency benchmark
Proceedings of the 2007 ACM SIGMOD international conference on Management of data
PARAID: A gear-shifting power-aware RAID
ACM Transactions on Storage (TOS)
Pergamum: replacing tape with energy efficient, reliable, disk-based archival storage
FAST'08 Proceedings of the 6th USENIX Conference on File and Storage Technologies
Write off-loading: practical power management for enterprise storage
FAST'08 Proceedings of the 6th USENIX Conference on File and Storage Technologies
SEA: A Striping-Based Energy-Aware Strategy for Data Placement in RAID-Structured Storage Systems
IEEE Transactions on Computers
Proceedings of the VLDB Endowment
Workload decomposition for power efficient storage systems
HotPower'08 Proceedings of the 2008 conference on Power aware computing and systems
IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Dynamic data reorganization for energy savings in disk storage systems
SSDBM'10 Proceedings of the 22nd international conference on Scientific and statistical database management
Hi-index | 0.00 |
Scientific data centers comprised of high-powered computing equipment and large capacity disk storage systems consume considerable amount of energy. Dynamic power management techniques (DPM) are commonly used for saving energy in disk systems. These involve powering down disks that exhibit long idle periods and placing them in standby mode. A file request from a disk in standby mode will incur both energy and performance penalties as it takes energy (and time) to spin up the disk before it can serve a file. For this reason, DPM has to make decisions as to when to transition the disk into standby mode such that the energy saved is greater than the energy needed to spin it up again and the performance penalty is tolerable. The length of the idle period until the DPM decides to power down a disk is called idleness threshold. In this paper, we study both analytically and experimentally dynamic power management techniques that save energy subject to performance constraints on file access costs. Based on observed workloads of scientific applications and disk characteristics, we provide a methodology for determining file assignment to disks and computing idleness thresholds that result in significant improvements to the energy saved by existing DPM solutions while meeting response time constraints. We validate our methods with simulations that use traces taken from scientific applications.