Scientific User Behavior and Data-Sharing Trends in a
Petascale File System
Event Type
Paper
State of the Practice
TimeWednesday, November 15th4:30pm -
5pm
Location405-406-407
DescriptionOak Ridge Leadership Computing Facility (OLCF) runs
some of the world's most powerful supercomputers,
connected through a center-wide petascale file system.
In this paper, we study the behavioral trends of 1,362
users and 387 projects from 39 scientific domains, by
analyzing daily file system snapshots collected over 500
days.
The scope of our analysis encompasses: (i) a quantitative system-centric metrics of the file system; (ii) user behavior on the file system; and (iii) data-sharing trends between users and projects. To the best of our knowledge, our work is the first of its kind to provide comprehensive insights on user behavior through file system snapshot analysis of a large-scale shared file system. We envision that this study will provide valuable insights on the design, operation, and management of storage systems at scale, and also encourage other HPC centers to undertake similar such efforts.
The scope of our analysis encompasses: (i) a quantitative system-centric metrics of the file system; (ii) user behavior on the file system; and (iii) data-sharing trends between users and projects. To the best of our knowledge, our work is the first of its kind to provide comprehensive insights on user behavior through file system snapshot analysis of a large-scale shared file system. We envision that this study will provide valuable insights on the design, operation, and management of storage systems at scale, and also encourage other HPC centers to undertake similar such efforts.
Download PDF:
here
