Understanding Data Access Patterns for dCache System
被引:0
|
作者:
Bellavita, Julian
论文数: 0引用数: 0
h-index: 0
机构:
Univ Calif Berkeley, Berkeley, CA 94720 USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Bellavita, Julian
[1
]
Sim, Caitlin
论文数: 0引用数: 0
h-index: 0
机构:
Univ Calif Berkeley, Berkeley, CA 94720 USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Sim, Caitlin
[1
]
Wu, Kesheng
论文数: 0引用数: 0
h-index: 0
机构:
Lawrence Berkeley Natl Lab, Berkeley, CA USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Wu, Kesheng
[2
]
Sim, Alex
论文数: 0引用数: 0
h-index: 0
机构:
Lawrence Berkeley Natl Lab, Berkeley, CA USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Sim, Alex
[2
]
Yoo, Shinjae
论文数: 0引用数: 0
h-index: 0
机构:
Brookhaven Natl Lab, Upton, NY USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Yoo, Shinjae
[3
]
Ito, Hiro
论文数: 0引用数: 0
h-index: 0
机构:
Brookhaven Natl Lab, Upton, NY USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Ito, Hiro
[3
]
Garonne, Vincent
论文数: 0引用数: 0
h-index: 0
机构:
Brookhaven Natl Lab, Upton, NY USAUniv Calif Berkeley, Berkeley, CA 94720 USA
Garonne, Vincent
[3
]
Lancon, Eric
论文数: 0引用数: 0
h-index: 0
机构:Univ Calif Berkeley, Berkeley, CA 94720 USA
Lancon, Eric
机构:
[1] Univ Calif Berkeley, Berkeley, CA 94720 USA
[2] Lawrence Berkeley Natl Lab, Berkeley, CA USA
[3] Brookhaven Natl Lab, Upton, NY USA
来源:
26TH INTERNATIONAL CONFERENCE ON COMPUTING IN HIGH ENERGY AND NUCLEAR PHYSICS, CHEP 2023
|
2024年
/
295卷
关键词:
D O I:
10.1051/epjconf/202429501053
中图分类号:
TP39 [计算机的应用];
学科分类号:
081203 ;
0835 ;
摘要:
The storage management system dCache acts as a disk cache for high-energy physics (HEP) data from the US ATLAS community. Since its disk capacity is considerably smaller than the total volume of ATLAS data, a heuristic is needed to determine what data should be kept on disks. An effective heuristic would be to keep the data files that are expected to be heavily accessed in the near future. Through a careful study of access statistics, we find a few most popular datasets are accessed much more frequently than others, even though these popular datasets change over time. If we could predict the near-term popularity of datasets, we could pin the most popular ones in the disk cache to prevent their accidental removal and guarantee their availability. To predict a dataset popularity, we present several methods for forecasting the number of times a dataset will be accessed in the next day. Test results show that these methods could predict the next-day access counts of popular datasets reliably. This observation is confirmed with dCache logs from two separate time ranges.