Publications
Detailed Information
Optimizing File Systems for High-Performance Storage Devices
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 염헌영 | - |
dc.contributor.author | 손용석 | - |
dc.date.accessioned | 2018-05-28T16:23:23Z | - |
dc.date.available | 2018-05-28T16:23:23Z | - |
dc.date.issued | 2018-02 | - |
dc.identifier.other | 000000149315 | - |
dc.identifier.uri | https://hdl.handle.net/10371/140691 | - |
dc.description | 학위논문 (박사)-- 서울대학교 대학원 : 공과대학 전기·컴퓨터공학부, 2018. 2. 염헌영. | - |
dc.description.abstract | High-performance storage technologies such as solid-state drives (SSDs) provide low-latency, high throughput, and high I/O parallelism to legacy storage systems. SSDs access data without mechanical overhead, and they often leads to order-of-magnitude improvements in performance over legacy storage devices such as hard disk drives (HDDs).
However, replacing HDDs with SSDs while keeping the software I/O stack or not exploiting SSD features does not lead to maximum performance. In this dissertation, we optimize file systems to fully exploit the SSD features (e.g., low-latency and high I/O parallelism). First, we analyze and explore I/O strategies in the existing file systems on low-latency SSDs. The file systems issue and complete several I/O requests when blocks are not contiguous, which does not take advantage of the low-latency of SSDs. To address this problem, we propose efficient I/O strategies, which transfer requests from discontiguous host memory buffers in the file systems to discontiguous storage segments in a single I/O request. Thus, they enable file systems to fully exploit the performance of low-latency SSDs. Second, we investigate the locking and I/O parallelism in the existing file systems on highly parallel SSDs. In the file systems, the coarse-grained locking to access shared data structures is used and I/O operations are serialized by a single thread. For these reasons, the file systems often face the problem of lock contention and underutilization of I/O bandwidth on multi-cores with highly parallel SSDs. To address these issues, we enable concurrent updates on data structures and parallelize I/O operations. We implement our techniques in EXT4/JBD2 and evaluate them on low-latency and highly parallel SSDs. The experimental results show that our optimized file system improves the performance compared to the existing EXT4 file system. | - |
dc.description.tableofcontents | Chapter 1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Approach and Contributions . . . . . . . . . . . . . . . . . . . . 3 1.3 Dissertation Structure . . . . . . . . . . . . . . . . . . . . . . . . 4 Chapter 2 Background 6 2.1 High-performance Storage Devices . . . . . . . . . . . . . . . . . 6 2.2 Crash Consistency in File Systems . . . . . . . . . . . . . . . . . 7 2.3 Read and Write Operations in the Existing File Systems . . . . . 9 2.4 Journal I/O in the Journaling File Systems . . . . . . . . . . . . 10 2.5 Recovery in the Journaling File Systems . . . . . . . . . . . . . . 13 2.6 Existing Locking and I/O Parallelism in Journaling File Systems 14 Chapter 3 Design and Implementation 24 3.1 Optimizing File Systems for Low-latency Storage Devices . . . . 24 3.1.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.1.2 Implementation . . . . . . . . . . . . . . . . . . . . . . . . 30 3.2 Optimizing File Systems for Highly Parallel Storage Devices . . . 33 3.2.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.2.2 Implementation . . . . . . . . . . . . . . . . . . . . . . . . 39 Chapter 4 Evaluation 50 4.1 Evaluating the Optimized File System for Low-latency Storage . 50 4.1.1 Run-time Performance . . . . . . . . . . . . . . . . . . . . 52 4.1.2 Recovery Performance . . . . . . . . . . . . . . . . . . . . 57 4.1.3 Experimental Analysis . . . . . . . . . . . . . . . . . . . . 59 4.2 Evaluating the Optimized File System for Highly Parallel Storage 61 4.2.1 Run-time Performance . . . . . . . . . . . . . . . . . . . . 63 4.2.2 Recovery Performance . . . . . . . . . . . . . . . . . . . . 66 4.2.3 Experimental Analysis . . . . . . . . . . . . . . . . . . . . 67 Chapter 5 Related Work 69 5.1 Analysis and Evaluation of High-Performance storage . . . . . . 69 5.2 Study of Journaling File Systems . . . . . . . . . . . . . . . . . . 70 5.3 File and I/O System Optimizations for Low-latency Storage . . . 72 5.4 Study of Scalability in Operating Systems . . . . . . . . . . . . . 75 5.5 File and I/O System Optimizations for Highly Parallel Storage . 75 Chapter 6 Conculsion 78 6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 6.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 | - |
dc.format | application/pdf | - |
dc.format.extent | 17623349 bytes | - |
dc.format.medium | application/pdf | - |
dc.language.iso | en | - |
dc.publisher | 서울대학교 대학원 | - |
dc.subject | File system | - |
dc.subject | Operating system | - |
dc.subject | High-Performance Storage Devices | - |
dc.subject | Solid-State Drive | - |
dc.subject.ddc | 621.3 | - |
dc.title | Optimizing File Systems for High-Performance Storage Devices | - |
dc.type | Thesis | - |
dc.description.degree | Doctor | - |
dc.contributor.affiliation | 공과대학 전기·컴퓨터공학부 | - |
dc.date.awarded | 2018-02 | - |
- Appears in Collections:
- Files in This Item:
Item View & Download Count
Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.