Current Result Document : 17 / 17
| 한글제목(Korean Title) |
Petaflop 슈퍼컴퓨터 스토리지의 데이터 중복 잠재성 분석 연구 |
| 영문제목(English Title) |
An Analysis of Data Deduplication Potential on Storage System of Petaflop Supercomputer |
| 저자(Author) |
조용현
사프다 자밀
김기현
이재국
안도식
홍태영
김영재
Yonghyeon Cho
Safdar Jamil
Kihyun Kim
Jae-Kook Lee
Dosik Ahn
Taeyoung Hong
Youngjae Kim
|
| 원문수록처(Citation) |
VOL 49 NO. 02 PP. 0051 ~ 0053 (2022. 12) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
The petaflop supercomputers have provided optimal resources for numerous scientific domains. Specifically, the storage nodes in supercomputers support tens to hundreds of petabytes of capacity and they are managed by parallel file system. Several studies have explored storage space optimization techniques, such as data deduplication (dedup). The dedup reduces the storage space by identifying and eliminating the duplicate data. However, these studies only considered the supercomputers at tera-scale. Therefore, in this work, we perform data dedup analysis on the Nurion petaflop supercomputer operated by Korea Institute of Science and Technology (KISTI) using an in-house dedup tool, named Deduplication Analyzer (Danzer). We observe that on average 25% of the data can be removed by adopting data dedup. |
| 키워드(Keyword) |
|
| 파일첨부 |
PDF 다운로드
|
