오픈 소스 기반 데이터 분산 중복 제거 파일 시스템의 성능 분석

다국어 초록 (Multilingual Abstract)

ZFS, LessFS, SDFS is deduplication file system. Deduplication file system performance comparison results. Lessfs were shown the execution time and the CPU utilization good results. SDFS were shown reduce Storage usage 1/8 than the general file system
In this paper, new system is propose that the SDFS and the Lessfs Combine Advantage. New system is use the two DFE and one DSE.
New system combines one DSE with two DFE and it is keep up to data integrity and consistency. Test environment make a comparison between Single DFE and Dual DFE.
Test result, Dual DFE was good better than Single DFE. Dual DFE target was reduced CPU usage and a fast deduplication time could confirm the result. In this paper, new system was achievement of the purposes.
Thereby new system can be solved that problem of large data storage increase and of power consumption.

목차 (Table of Contents)

• 1. 서론 5
• 1.1. 연구 배경 및 목적 5
• 1.2. 연구 내용 및 범위 6
• 1.3. 논문 구성 8
• 2. 관련연구 10
• 2.1. 중복제거와 중복제거처리방식 10
• 2.2. Open Source Based De-duplication 기술 16
• 2.2.1. ZFS(The Z File System) 16
• 2.2.2. Lessfs(Data De-duplication for Less) 17
• 2.2.3. SDFS(A user space De-duplication File System) 18
• 2.3. Open Source Based De-duplication 장단점 19
• 3. 분산 중복제거 하드웨어 구조 21
• 3.1. 서버 구조 22
• 3.2. 스토리지 구조 23
• 3.2.1. 스토리지의 개요 23
• 3.2.2. 스토리지의 구조 23
• 3.3. SAN/NAS/DAS의 구조 24
• 3.3.1. SAN(Storage Attached Network) 24
• 3.3.2. NAS(Network Attached Storage) 26
• 3.3.3. DAS(Direct Attached Storage) 28
• 4. 분산 중복제거 소프트웨어 구조 29
• 4.1. SDFS(A User Space De-duplication File System) 30
• 4.2. DFE(De-duplication File Engine) 32
• 4.3. DSE(De-duplication File Engine) 33
• 4.4. Dual DFE(De-duplication File Engine) 35
• 4.5. GFS2(Global File System) 36
• 4.6. Fuse(File System in User Space) 38
• 5. 성능 분석 39
• 5.1. 성능 분석 시나리오 39
• 5.1.1. Single DFE 데이터 입력 방식 39
• 5.1.2. Dual DFE 데이터 입력 방식 40
• 5.1.3. Dual DFE Meta 동기화 42
• 5.1.4. 성능비교 시나리오 43
• 5.2. 제안 시스템 성능 분석 45
• 5.2.1. 가상화 환경 Dual DFE 성능 분석 45
• 5.2.2. 물리적 환경 Single DFE와 Dual DFE의 압축 파일 분산 중복제거 성능 분석 49
• 5.2.3. 물리적 환경 Single DFE와 Dual DFE의 일반 파일 분산 중복제거 성능 분석 51
• 6. 결론 55
• 참고문헌 56