- 學位論文
Ceph軟體定義分散式儲存系統效能優化之研究
The Study of Performance Optimization for the Ceph Software Defined Storage System
摘要
傳統的儲存服務架構受限於不同儲存設備廠商所提供儲存設備與技術，導致擴充及建置上不夠靈活，不能有效控管資源。因此，在此狀況之下，許多企業與廠商紛紛提出透過虛擬化、軟體定義的方式來解決此問題，在軟體定義儲存上以Ceph分散式儲存系統最受關注，Ceph是一種統一式儲存系統，並且被設計為可水平擴展且無單點故障的儲存系統，而在Ceph中提供多種軟體功能，因此可透過使用Ceph分散式儲存系統壓縮功能來有效降低儲存成本並增加整體儲存叢集效能，而在壓縮功能上則可透過設定壓縮壓縮等級與參數來達到不同的壓縮率與效能，但這些參數若隨意調整會對於儲存系統整體效能造成影響。因此，本研究基於Ceph Plugin 架構結合平行化壓縮演算法來降低物件式儲存成本，並且透過效能基準測試基準工具，來評估Ceph儲存叢集與分析最佳的參數設定與優化，以快速提供Ceph儲存叢集建置者選擇符合建置環境需求的參數設定。
並列摘要
Traditional storage service architectures are not enough scalability and flexibility, resulting in the inability to effectively utilize storage resources. Therefore, many enterprises and vendors have proposed to solve this problem through virtualization and software-defined technologies. And in the software defined storage solution, Ceph is the popular open source software solution. Ceph is a unified storage system, and it is designed for storage systems that scale horizontally and have no single point of failure. Also, Ceph offers a variety of software functions, so it can use compression function to reduce storage costs and increase storage clusters performance, and then can be achieved different compression ratio and performance through the setting compression level and feature parameters. However, if these parameters are adjusted at will, it will affect the overall performance of the storage system. Therefore, this study is based on the Ceph Plugin architecture combined with the parallel compression algorithm to reduce the storage and storage costs of the object storage and evaluate the optimal setting through the performance benchmark tool analyzing. Then, quickly provide optimized settings for Ceph system based on user requirement.
參考文獻
延伸閱讀
- 洪士哲（2016）。基於Ceph分散式儲存上資料可靠性之效能評測〔碩士論文，國立臺中科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0061-2907201616233900
- 林春旺（2019）。基於Ceph軟體定義分散式儲存架構設計私有區塊鏈之儲存優化〔碩士論文，國立臺中科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0061-2108201913521900
- 林仕軒（2016）。透過有效率之叢集Sufferage排程演算法來提升異質雲端環境之系統執行效能〔碩士論文，朝陽科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0078-1108201714030876
- Kuo, K. T. (2014). 分區定址動態繞徑演算法的開發與實踐在軟體定義網路之資料中心 [master's thesis, National Chiao Tung University]. Airiti Library. https://doi.org/10.6842/NCTU.2014.00866
- Huang, C. W. (2014). Performance and Energy Optimization of Embedded Memory Systems [doctoral dissertation, National Chiao Tung University]. Airiti Library. https://doi.org/10.6842/NCTU.2014.01110