發(fā)布時間：2018-07-21 08:58

【摘要】：隨著對存儲系統(tǒng)容量的要求不斷增大，海量存儲系統(tǒng)已成為趨勢。存儲系統(tǒng)評測是存儲系統(tǒng)研究的一個重要方面，是存儲系統(tǒng)研究與應用的重要理論基礎和支撐技術(shù)。伴隨著存儲系統(tǒng)的發(fā)展，用于系統(tǒng)評測的模型、方法和工具也在不斷地更新。 本文對海量存儲系統(tǒng)進行了全面的研究，包括其發(fā)展過程、各種技術(shù)的詳細內(nèi)容。海量存儲的提出與發(fā)展伴隨著人們對數(shù)據(jù)存儲需求的不斷提高，從直連式的存儲設備到存儲網(wǎng)絡的成熟，從單機式文件系統(tǒng)到分布式文件系統(tǒng)的廣泛流行，從SCSI標準協(xié)議到FC、iSCSI等擴展協(xié)議的普及，從幾G的磁盤容量到幾TB甚至幾PB的存儲陣列問世，，存儲系統(tǒng)的發(fā)展經(jīng)歷了翻天覆地的變化。文中從海量存儲系統(tǒng)典型結(jié)構(gòu)的四個層次——應用層、文件系統(tǒng)層、存儲網(wǎng)絡層和存儲設備層分別進行了容錯機制剖析，并針對多種流行的技術(shù)提出了相應的容錯能力測試策略。 針對海量存儲系統(tǒng)的分層結(jié)構(gòu)，本文設計并實現(xiàn)了一個海量存儲系統(tǒng)容錯能力測試平臺，根據(jù)上述研究的容錯能力測試策略在多個層次設計并實現(xiàn)了容錯能力測試套件，還根據(jù)海量存儲系統(tǒng)測試的需要設計與實現(xiàn)了多個負載發(fā)后套件與監(jiān)測套件，并將這些套件集成于測試平臺上。平臺將可以結(jié)合相應的套件通過軟件故障注入的方式對目標海量存儲系統(tǒng)進行測試，并通過收集數(shù)據(jù)等方式為容錯能力評價提供支持。 最后，本文利用所開發(fā)的平臺對特定的目標系統(tǒng)進行了容錯能力測試，對平臺的功能進行了驗證。
[Abstract]:With the increasing demand for storage system capacity, mass storage system has become a trend. Storage system evaluation is an important aspect of storage system research. It is an important theoretical basis and supporting technology for storage system research and application. With the development of storage system, the models, methods and tools for system evaluation are constantly updated. This paper makes a comprehensive study on mass storage system, including its development process, various technology details. With the development and development of mass storage, with the increasing demand for data storage, from direct-connected storage devices to the maturity of storage network, from single-machine file system to distributed file system is widely popular. From SCSI standard protocol to FCU iSCSI extension protocol, from several gigabytes of disk capacity to several terabytes and even several PB storage arrays, the development of storage system has experienced earth-shaking changes. In this paper, the fault-tolerant mechanism of the typical structure of mass storage system is analyzed from four layers: application layer, file system layer, storage network layer and storage device layer. At the same time, the paper puts forward the corresponding fault-tolerant ability test strategy for various popular technologies. Aiming at the hierarchical structure of mass storage system, this paper designs and implements a fault tolerance test platform of mass storage system. According to the above research fault-tolerant ability test strategy, we design and implement the fault tolerance test suite at multiple levels. According to the needs of mass storage system testing, several post-load packages and monitoring suites are designed and implemented, and these suites are integrated into the test platform. The platform will be able to test the target mass storage system through software fault injection combined with the corresponding suite, and provide support for fault tolerance evaluation by collecting data. Finally, the fault-tolerant ability of a specific target system is tested using the developed platform, and the function of the platform is verified.
【學位授予單位】：哈爾濱工業(yè)大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：TP333

【參考文獻】

相關(guān)期刊論文 前6條

1 王閻;康慕寧;張曉;胡娜;;海量存儲系統(tǒng)評測體系的研究[J];電子設計工程;2011年07期

2 趙曉南;李戰(zhàn)懷;張曉;曾雷杰;;基于trace的網(wǎng)絡存儲系統(tǒng)評測研究[J];計算機科學;2010年02期

3 張媛;盧澤新;劉亞萍;;NFS over Lustre性能評測與分析[J];計算機工程;2007年10期

4 李鑫;李戰(zhàn)懷;張曉;;面向海量存儲的文件系統(tǒng)評測基準[J];計算機工程;2011年15期

5 譚毓安;信息存儲新技術(shù):NAS和SAN[J];情報理論與實踐;2002年04期

6 侯f ;趙躍龍;王禹;;海量存儲系統(tǒng)的層次化體系結(jié)構(gòu)[J];小型微型計算機系統(tǒng);2010年06期

本文編號：2135009