軟件失效模式評測工具的設(shè)計與實現(xiàn)
發(fā)布時間:2019-06-14 15:39
【摘要】:計算機系統(tǒng)已經(jīng)廣泛應(yīng)用于各個領(lǐng)域,用戶在使用計算機系統(tǒng)提供的強大服務(wù)的同時也非常關(guān)心軟件系統(tǒng)本身的容錯性能,,在遇到系統(tǒng)故障時,是否仍然能夠正常的執(zhí)行并使用戶得到正確的服務(wù)。然而,對軟件容錯性的評測仍然存在著評測局限性大、數(shù)據(jù)收集效率低等問題,需要新的方法對軟件的容錯性進行更為規(guī)范的評價。 故障注入是對容錯性進行評測的常用方法,也是很有效的方法。本課題通過對典型FARM故障注入模型的研究,提出了針對軟件失效影響研究的改進的故障注入模型,并確定了模型中各集合的內(nèi)容。以基于調(diào)試器的故障注入手段為基礎(chǔ),提出了軟件失效影響的評測思路,并對失效影響評測的策略進行了說明。 在故障注入模型的指導(dǎo)下,本文通過使用調(diào)試器的故障注入方法以及系統(tǒng)進程運行狀態(tài)的監(jiān)控方法,實現(xiàn)了一個以軟件失效影響為研究目標,在windowsNT系統(tǒng)下可以進行批量的非永久性故障注入的失效影響測試平臺,建立了軟件失效影響的測試環(huán)境。 為了驗證本文所實現(xiàn)的失效影響評測工具的有效性以及對軟件的失效影響進行初步的研究,本文選擇了一些有代表性的基準程序進行了各類型的軟件失效影響評測,并對已有的軟件容錯方法進行了評測。實驗結(jié)果表明了本文所設(shè)計實現(xiàn)的評測工具的有效性,對軟件的失效影響進行了初步的分析,并且通過對不同組測試結(jié)果的分析,得到了軟件失效的一些規(guī)律,并對已有軟件容錯方法的性能做了驗證。
[Abstract]:The computer system has been widely used in various fields, and the user is very concerned with the fault-tolerance of the software system itself while using the powerful service provided by the computer system. However, the evaluation of the fault-tolerance of the software still has the problems of large evaluation limitation, low data collection efficiency and the like, and a new method is needed to evaluate the fault-tolerance of the software more standard. Fault injection is a common method for evaluating fault-tolerance, and it is also a very effective method. In this paper, through the study of typical FARM fault injection model, an improved fault injection model for the study of software failure impact is proposed, and the interior of each set in the model is determined. In this paper, based on the fault injection method based on the debugger, the evaluation method of the software failure effect is put forward, and the strategy of the failure impact evaluation is described. In the guidance of the fault injection model, the fault injection method of the debugger and the monitoring method of the running state of the system process are used in this paper. In this paper, a test platform for failure of a batch of non-permanent fault injection can be carried out under the window NT system, and the effect of software failure is established. In order to verify the effectiveness of the failure effect evaluation tool and the effect of the failure of the software, some representative benchmark programs have been selected to evaluate the effect of the software failure, and the existing software fault-tolerant method is given. The evaluation is carried out. The experimental results show that the effectiveness of the evaluation tool designed in this paper is a preliminary analysis of the effect of the software, and some rules of the software failure are obtained through the analysis of the test results of different groups, and the performance of the fault-tolerant method of the existing software is also analyzed.
【學(xué)位授予單位】:哈爾濱工業(yè)大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2012
【分類號】:TP302.8
本文編號:2499496
[Abstract]:The computer system has been widely used in various fields, and the user is very concerned with the fault-tolerance of the software system itself while using the powerful service provided by the computer system. However, the evaluation of the fault-tolerance of the software still has the problems of large evaluation limitation, low data collection efficiency and the like, and a new method is needed to evaluate the fault-tolerance of the software more standard. Fault injection is a common method for evaluating fault-tolerance, and it is also a very effective method. In this paper, through the study of typical FARM fault injection model, an improved fault injection model for the study of software failure impact is proposed, and the interior of each set in the model is determined. In this paper, based on the fault injection method based on the debugger, the evaluation method of the software failure effect is put forward, and the strategy of the failure impact evaluation is described. In the guidance of the fault injection model, the fault injection method of the debugger and the monitoring method of the running state of the system process are used in this paper. In this paper, a test platform for failure of a batch of non-permanent fault injection can be carried out under the window NT system, and the effect of software failure is established. In order to verify the effectiveness of the failure effect evaluation tool and the effect of the failure of the software, some representative benchmark programs have been selected to evaluate the effect of the software failure, and the existing software fault-tolerant method is given. The evaluation is carried out. The experimental results show that the effectiveness of the evaluation tool designed in this paper is a preliminary analysis of the effect of the software, and some rules of the software failure are obtained through the analysis of the test results of different groups, and the performance of the fault-tolerant method of the existing software is also analyzed.
【學(xué)位授予單位】:哈爾濱工業(yè)大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2012
【分類號】:TP302.8
【參考文獻】
相關(guān)期刊論文 前4條
1 譚玲,曲峰,董劍,楊孝宗;基于軟件故障注入的容錯性能評測技術(shù)[J];計算機工程與科學(xué);2005年11期
2 王睿伯;盧錫城;盧凱;王紹剛;;面向CC-NUMA體系結(jié)構(gòu)的事務(wù)內(nèi)存沖突規(guī)避方法[J];計算機學(xué)報;2011年04期
3 袁靜;胡昌華;徐瑞;龍勇;;一類改進故障樹分析法的可靠性仿真及應(yīng)用[J];計算機應(yīng)用研究;2006年11期
4 溫東新,劉宏偉,楊孝宗,董劍;利用軟件故障注入提高軟件測試覆蓋率[J];同濟大學(xué)學(xué)報(自然科學(xué)版);2002年10期
本文編號:2499496
本文鏈接:http://sikaile.net/kejilunwen/jisuanjikexuelunwen/2499496.html
最近更新
教材專著