2010
DOI: 10.1145/1837915.1837918
|View full text |Cite
|
Sign up to set email alerts
|

Optimizing energy and performance for server-class file system workloads

Abstract: of the ThesisOptimizing Energy and Performance for Server-Class File System Workloads by Priya Sehgal Master of Science in Computer ScienceStony Brook University 2010Recently, power has emerged as a critical factor in designing components of storage systems, especially for power-hungry data centers. While there is some research into power-aware storage stack components, there are no systematic studies evaluating each component's impact separately. Various factors like workloads, hardware configurations, and so… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
10
0

Year Published

2011
2011
2016
2016

Publication Types

Select...
5
2

Relationship

1
6

Authors

Journals

citations
Cited by 13 publications
(12 citation statements)
references
References 28 publications
0
10
0
Order By: Relevance
“…Fei et al carried out a study of energy-saving/performance trade-offs in the context of Grid resource management [21], by applying and extending the theory of M/G/1 queues with vacations. Factors influencing energy and performance in data center storage systems (including deployed file systems) were studied in [17] while the disruptive impact of SSD technology on the energy and performance required to process database queries is considered in [16]. It was concluded that it is necessary to completely reassess traditional rules of thumb related to circumstances under which it is appropriate to use different kinds of table scans (i.e.…”
Section: Related Workmentioning
confidence: 99%
“…Fei et al carried out a study of energy-saving/performance trade-offs in the context of Grid resource management [21], by applying and extending the theory of M/G/1 queues with vacations. Factors influencing energy and performance in data center storage systems (including deployed file systems) were studied in [17] while the disruptive impact of SSD technology on the energy and performance required to process database queries is considered in [16]. It was concluded that it is necessary to completely reassess traditional rules of thumb related to circumstances under which it is appropriate to use different kinds of table scans (i.e.…”
Section: Related Workmentioning
confidence: 99%
“…Measuring the effect of changes to individual system parameters (such as the bandwidth of the disk controllers or the number of disks in the system) within this consistent framework enables the effect of these parameters within a complex system to be better understood. However, often only a fixed set of components are used (see, for example, [13], [17]) and few studies have measured how individual parameter changes affect overall system metrics. Without rigorous benchmarking across a range of possibilities for a single system parameter only rough estimates of the change in system metrics can be made and these estimates are highly reliable on vendor specifications.…”
Section: Measurement Of Archival Storagementioning
confidence: 99%
“…Studies in [6], [14], [18], [23] examined various optimal disk spin up/down time-out selections, data flushing, file-grouping and reliability to maximize the energy conservation while minimizing the impact on performance. [17] measured how different file systems (e.g., ext2/3, reserfs and xfs) have performance and power consumptions on two servers, finding that different file systems can be suited for different workloads. A system-wide energy consumption model in [13] measures individual components of the CPU, DRAM, HDD, fan and system board, to combine them for benchmarking and predicting overall energy consumption.…”
Section: Related Workmentioning
confidence: 99%
“…Most of the discussions about tail packing focused on the performance evaluation [16,18]; Sehgal et al discussed the influence of power consumption in a storage device [17], and Lu et al presented the impact of the lifetime of flash-based storage caused by the tailpacking solution [11] . Only a few prior works proposed solutions for increasing space utilization in file systems [4,14,15,21].…”
Section: Introductionmentioning
confidence: 99%