用数学建模方法评价存储系统性能
|
摘要
作为信息化时代的重要资源，存储系统具有重大的实用价值。随着数字信息的爆炸性增长，大数量的用户群和种类繁多的应用导致了人们对大容量信息存储系统的需求。为了方便许多资源的使用，发展存储系统是首选的普遍的途径，所以存储系统是应用广泛的，并且在迅速地发展。
在当今存储系统中，由于I/O性能是存储系统性能不可或缺的一部分，因此如何构造一个高性能、低能耗、适合不同应用的I/O系统就成为了一个重要的任务。存储系统的性能无论是对存储系统自身还是基于存储系统的应用都是至关重要的。其中I/O性能是评价存储系统的重要指标，对存储系统I/O性能的研究具有很大的理论和实用价值。目前计算机的I/O性能已经成为严重制约系统整体性能的瓶颈，所以I/O性能的分析与优化方法对于提升I/O性能就显得尤为重要。
近年来，排队论方法已经成为分析和探讨存储系统性能的主要方法，因此本论文以排队论为主要工具来分析和研究有关存储系统I/O方面的性能。如今，在大规模存储系统中，由于数据的增长和大量存储设备的存在，使得几乎每天都会有存储设备发生故障，因此，存储设备的可修性也成为一个必须考虑的因素。影响存储系统性能的因素很多，本论文分析和研究的重点是前端排队方式对整个存储系统的影响。已有很多专家分析和研究了存储系统的总体特征，因此，本论文主要分析一些局部特征，而且是针对存储系统I/O的问题建立相应的模型来分析I/O的服务设备、结构布置、运作模式等等。本文使用排队论和拟生灭过程相结合的方法分析了I/O性能，主要是通过排队论的一些指数分布和拟生灭过程的Q矩阵方法，例如假设服务效率等参数服从负指数分布，提出了矩阵模式的分析方法，总结了I/O运行的具体相关性能指标表达式。然后进行仿真，经过多次仿真得到参数之间的规律关系，将用来仿真的数据用于I/O具体的操作过程中，最后总结I/O不同情况的操作规律以及所有运作模式的普遍规律，所得结果为I/O性能评价以及调度策略的确定提供了的依据。
缓存是影响存储系统性能的重要因素。本文提出一种符合局部性原理的访问序列生成过程。这个过程具有局部性、延续性和突变性等良好性质，并且可以通过几个参数的设置控制以上各性质的强弱，生成符合各种访问模式的数字流，并对比各种缓存替换算法的优劣。在此方法的基础上，针对应用服务器和存储服务器构成的一个二级buffercache结构进行了研究。现有的ULC（Unified Level-Aware Caching）机制可有效解决多级缓存的数据冗余及在存储服务器端缓存的访问弱局部性问题。但是当存储服务器连接多个应用服务器时，ULC采用LRU策略为各应用服务器在存储服务器端分配缓存容量，该方法不能使存储服务器端缓存资源的边际收益最大化，为此本文提出一种多应用共享缓存的二级buffer cache动态分配策略MG-ULC（Marginal Gain-based ULC）。该策略以ULC机制为基础，根据各应用的访问模式在二级buffer cache的边际增益动态分配缓存容量。实验结果显示，随着各应用服务器访问模式的变化，MG-ULC能比ULC更合理地分配二级缓存，从而实现更高的缓存利用率，使存储系统性能进一步提升。
I/O负载是影响存储系统性能的重要因素，响应时间是反映存储系统性能的重要指标。为了提高存储系统的性能，本文对I/O负载的特征参数和响应时间进行了研究分析。为实现根据I/O负载的一些特征参数预测系统响应时间的目的，本文利用灰色系统理论和BP人工神经网络相结合的方法建立模型，确定了特征参数和响应时间之间的非线性关系，利用disksim得出trace数据，进行仿真，并将预测结果和单纯使用灰色和神经网路预测的结果进行比较，突出其优越性。由于考虑的因素不同，本文又结合BP神经网络和马尔科夫链，建立了一种新的用于I/O负载的预测模型BP神经网络-马尔科夫链预测模型（BP-MC）。通过对训练样本的学习，利用BP神经网络实现了对负载时间序列的滚动预测，同时得到了实测值与预测值的相对误差。在此基础上利用马尔科夫链对相对误差进行修正，有效的提高了预测结果的精度。将该模型应用于存储系统I/O负载预测中，结果表明该模型预测精度高，为存储系统性能预测提供了新的途径。
I/O性能评价在存储系统的设计和应用中占有非常重要的位置，如何快速有效地评价I/O的性能是优化I/O性能的一个重要步骤。本文提出了用定性和定量相结合的方法，即层次分析方法和人工神经网络方法相结合的方法来分析和研究I/O的性能，同样用disksim得出的trace数据进行仿真，仿真结果显示了其优越性，通过实例说明了此评价方法可以有效地评价存储系统的I/O性能。
As an important resource, the storage system is of great significance. With the explosivegrowth of digital information, users and various applications, the demands of high-capacitystorage systems is increasing very fast. The development of the storage system is thepreferred and prevalent way for the good use of many resources.
In today’s storage system, the performance of I/O system is an essential part of theperformance of storage system. Therefore, how to construct an I/O system with highperformance, low power consumption and suitable for different applications becomes animportant task. The performance of storage system is vital to the storage system itself and theapplication of storage systems. I/O performance is one of the most important performanceindicators for evaluating the storage system, and it has great theoretical and practical value forthe study of I/O performance of storage system. At present, because I/O performance ofcomputer has already become the bottleneck of the whole system for a long time, I/Operformance analysis and optimization method looks especially important and valuable.
In recent years, as the main methods of storage system performance, the queuing theorymethod has been analyzed and discussed. Therefore, in the thesis, the queuing theory is takenas the main tool to analyzes and study the performance of I/O aspects of storage system.Because of the growth of information and the existence of many storage devices in thelarge-scale storage nowadays, the storage device failures can occur almost every day, thus, therepairability of storage devices must to be a considerable factor. There are too many factorsinfluencing storage system performance. The paper focuses on the front-end queue of thewhole storage system.This thesis mainly analyzes some local characteristics, and aiming atthe I/O problem of storage system, the service equipment, structural arrangement, andoperation model of the I/O and so on are analyzed by establishing corresponding model. Inthe thesis, I/O performance model is analyzed by the use of the method of the combination ofquasi birth and death process and queuing model, and it is studied mainly passing theindicator distribution of queuing theory and the Q matrix pattern of quasi birth and deathprocess, such as service efficiency submit to megative exponential distribution and so on. Theanalytical method of matrix pattern is presented, and the I/O operational specific performanceindicator expression is summed up. Then after many simulations, the regular pattern isobtained. The data are use for the I/O concrete operation process, finally, the I/O operation rule under the different conditions and all the universal law under the operation mode aresummarized. The results may be used to evaluate storage performance, and are the basis fordeciding I/O scheduling strategy.
Cache is one of the important factors affect the storage system performance. This paperproposes a visit sequence generation process comply with the principle of locality. Theprocess has locality, continuity and mutability properties and so on, and can control thestrength of the above properties by a few parameters set. The digital flow is generated in linewith the various access patterns, and compared with various cache replacement algorithm. Onthe basis of this method, a two-level buffer cache structure that is constituted by the buffercaches of application servers and storage servers is studied. The existing ULC (UnifiedLevel-Aware Caching) protocol can effectively solve the problems that redundantly cachedblocks in multilevel hierarchy and weakly localized at storage server cache. However, whenthere are multiple application servers sharing one storage server, the ULC adopts LRUstrategy to allocate cache capacity of storage server to each application server, and thismethod can not gain the maximal marginal profits of the storage server cache. A second-levelbuffer cache dynamic allocation strategy called MG-ULC (Marginal Gain-based ULC) isproposed, and it is designed for storage servers in which multiple applications share the samecache resources. Based on the ULC protocol, the MG-ULC dynamically allocates cachecapacity in accordance with the second-level buffer cache marginal gain of each application.The results shows that, as each application's access pattern changes, the MG-ULC canallocate second-level buffer cache more rationally than the ULC, thereby realizing a highercache utilization.
I/O load is an important factor to affect the performance of storage system, and theresponse time is an important indicator to reflect the performance of storage system. In orderto improve the performance of storage system, the characteristic parameters of the I/O loadand response time are analyzed in this paper. In order to achieve the purpose of forecastingresponse time by some characteristic parameter of I/O load, the model is built by combiningwith the Grey System Theory and BP Artificial Neural Network method, and the nonlinearrelations of characteristic parameter and response time is determined. Make use of the disksim,the trace data is obtained, then the trace data is used in simulating. The predicted result iscompared with the results of using grey and neural network prediction, and the advantages arehighlighted. Due to the considered factors are different, a new prediction model is proposedby combining with the BP neural network and Markov Chain, named as BP neural network-Markov chain(BP-MC) model and applied to I/O load prediction. Through emulatingthe training sample, the rolling prediction of load time series is achieved by BP neuralnetwork, and the relative error of measured and predicted values is acquired. Applying theMarkov Chain to correct the relative error and the accruacy of predicting results is improvedeffectively. The prediction model was used to predict the I/O load of Storage system， theresult shows that the model is high-precision， so it provides a new approach for Storagesystem prediction.
The evaluation of I/O performance is very important in the design and use of storagesystem. How to evaluate the I/O performance quickly and efficiently has become an importantprocedure for optimizing I/O performance. In this thesis, both qualitative and quantitativemethod is proposed, which is the combination of analytical hierarchy process level analysismethod and artificial neural network method. I/O performance is analyzed and studied by thismethod. Similarly, the trace data used for simulation is obtained by making use of the disksim.The advantage is found from the simulate result. The evaluation methodology can effectivelyevaluate the I/O performance of storage system.
在当今存储系统中，由于I/O性能是存储系统性能不可或缺的一部分，因此如何构造一个高性能、低能耗、适合不同应用的I/O系统就成为了一个重要的任务。存储系统的性能无论是对存储系统自身还是基于存储系统的应用都是至关重要的。其中I/O性能是评价存储系统的重要指标，对存储系统I/O性能的研究具有很大的理论和实用价值。目前计算机的I/O性能已经成为严重制约系统整体性能的瓶颈，所以I/O性能的分析与优化方法对于提升I/O性能就显得尤为重要。
近年来，排队论方法已经成为分析和探讨存储系统性能的主要方法，因此本论文以排队论为主要工具来分析和研究有关存储系统I/O方面的性能。如今，在大规模存储系统中，由于数据的增长和大量存储设备的存在，使得几乎每天都会有存储设备发生故障，因此，存储设备的可修性也成为一个必须考虑的因素。影响存储系统性能的因素很多，本论文分析和研究的重点是前端排队方式对整个存储系统的影响。已有很多专家分析和研究了存储系统的总体特征，因此，本论文主要分析一些局部特征，而且是针对存储系统I/O的问题建立相应的模型来分析I/O的服务设备、结构布置、运作模式等等。本文使用排队论和拟生灭过程相结合的方法分析了I/O性能，主要是通过排队论的一些指数分布和拟生灭过程的Q矩阵方法，例如假设服务效率等参数服从负指数分布，提出了矩阵模式的分析方法，总结了I/O运行的具体相关性能指标表达式。然后进行仿真，经过多次仿真得到参数之间的规律关系，将用来仿真的数据用于I/O具体的操作过程中，最后总结I/O不同情况的操作规律以及所有运作模式的普遍规律，所得结果为I/O性能评价以及调度策略的确定提供了的依据。
缓存是影响存储系统性能的重要因素。本文提出一种符合局部性原理的访问序列生成过程。这个过程具有局部性、延续性和突变性等良好性质，并且可以通过几个参数的设置控制以上各性质的强弱，生成符合各种访问模式的数字流，并对比各种缓存替换算法的优劣。在此方法的基础上，针对应用服务器和存储服务器构成的一个二级buffercache结构进行了研究。现有的ULC（Unified Level-Aware Caching）机制可有效解决多级缓存的数据冗余及在存储服务器端缓存的访问弱局部性问题。但是当存储服务器连接多个应用服务器时，ULC采用LRU策略为各应用服务器在存储服务器端分配缓存容量，该方法不能使存储服务器端缓存资源的边际收益最大化，为此本文提出一种多应用共享缓存的二级buffer cache动态分配策略MG-ULC（Marginal Gain-based ULC）。该策略以ULC机制为基础，根据各应用的访问模式在二级buffer cache的边际增益动态分配缓存容量。实验结果显示，随着各应用服务器访问模式的变化，MG-ULC能比ULC更合理地分配二级缓存，从而实现更高的缓存利用率，使存储系统性能进一步提升。
I/O负载是影响存储系统性能的重要因素，响应时间是反映存储系统性能的重要指标。为了提高存储系统的性能，本文对I/O负载的特征参数和响应时间进行了研究分析。为实现根据I/O负载的一些特征参数预测系统响应时间的目的，本文利用灰色系统理论和BP人工神经网络相结合的方法建立模型，确定了特征参数和响应时间之间的非线性关系，利用disksim得出trace数据，进行仿真，并将预测结果和单纯使用灰色和神经网路预测的结果进行比较，突出其优越性。由于考虑的因素不同，本文又结合BP神经网络和马尔科夫链，建立了一种新的用于I/O负载的预测模型BP神经网络-马尔科夫链预测模型（BP-MC）。通过对训练样本的学习，利用BP神经网络实现了对负载时间序列的滚动预测，同时得到了实测值与预测值的相对误差。在此基础上利用马尔科夫链对相对误差进行修正，有效的提高了预测结果的精度。将该模型应用于存储系统I/O负载预测中，结果表明该模型预测精度高，为存储系统性能预测提供了新的途径。
I/O性能评价在存储系统的设计和应用中占有非常重要的位置，如何快速有效地评价I/O的性能是优化I/O性能的一个重要步骤。本文提出了用定性和定量相结合的方法，即层次分析方法和人工神经网络方法相结合的方法来分析和研究I/O的性能，同样用disksim得出的trace数据进行仿真，仿真结果显示了其优越性，通过实例说明了此评价方法可以有效地评价存储系统的I/O性能。
As an important resource, the storage system is of great significance. With the explosivegrowth of digital information, users and various applications, the demands of high-capacitystorage systems is increasing very fast. The development of the storage system is thepreferred and prevalent way for the good use of many resources.
In today’s storage system, the performance of I/O system is an essential part of theperformance of storage system. Therefore, how to construct an I/O system with highperformance, low power consumption and suitable for different applications becomes animportant task. The performance of storage system is vital to the storage system itself and theapplication of storage systems. I/O performance is one of the most important performanceindicators for evaluating the storage system, and it has great theoretical and practical value forthe study of I/O performance of storage system. At present, because I/O performance ofcomputer has already become the bottleneck of the whole system for a long time, I/Operformance analysis and optimization method looks especially important and valuable.
In recent years, as the main methods of storage system performance, the queuing theorymethod has been analyzed and discussed. Therefore, in the thesis, the queuing theory is takenas the main tool to analyzes and study the performance of I/O aspects of storage system.Because of the growth of information and the existence of many storage devices in thelarge-scale storage nowadays, the storage device failures can occur almost every day, thus, therepairability of storage devices must to be a considerable factor. There are too many factorsinfluencing storage system performance. The paper focuses on the front-end queue of thewhole storage system.This thesis mainly analyzes some local characteristics, and aiming atthe I/O problem of storage system, the service equipment, structural arrangement, andoperation model of the I/O and so on are analyzed by establishing corresponding model. Inthe thesis, I/O performance model is analyzed by the use of the method of the combination ofquasi birth and death process and queuing model, and it is studied mainly passing theindicator distribution of queuing theory and the Q matrix pattern of quasi birth and deathprocess, such as service efficiency submit to megative exponential distribution and so on. Theanalytical method of matrix pattern is presented, and the I/O operational specific performanceindicator expression is summed up. Then after many simulations, the regular pattern isobtained. The data are use for the I/O concrete operation process, finally, the I/O operation rule under the different conditions and all the universal law under the operation mode aresummarized. The results may be used to evaluate storage performance, and are the basis fordeciding I/O scheduling strategy.
Cache is one of the important factors affect the storage system performance. This paperproposes a visit sequence generation process comply with the principle of locality. Theprocess has locality, continuity and mutability properties and so on, and can control thestrength of the above properties by a few parameters set. The digital flow is generated in linewith the various access patterns, and compared with various cache replacement algorithm. Onthe basis of this method, a two-level buffer cache structure that is constituted by the buffercaches of application servers and storage servers is studied. The existing ULC (UnifiedLevel-Aware Caching) protocol can effectively solve the problems that redundantly cachedblocks in multilevel hierarchy and weakly localized at storage server cache. However, whenthere are multiple application servers sharing one storage server, the ULC adopts LRUstrategy to allocate cache capacity of storage server to each application server, and thismethod can not gain the maximal marginal profits of the storage server cache. A second-levelbuffer cache dynamic allocation strategy called MG-ULC (Marginal Gain-based ULC) isproposed, and it is designed for storage servers in which multiple applications share the samecache resources. Based on the ULC protocol, the MG-ULC dynamically allocates cachecapacity in accordance with the second-level buffer cache marginal gain of each application.The results shows that, as each application's access pattern changes, the MG-ULC canallocate second-level buffer cache more rationally than the ULC, thereby realizing a highercache utilization.
I/O load is an important factor to affect the performance of storage system, and theresponse time is an important indicator to reflect the performance of storage system. In orderto improve the performance of storage system, the characteristic parameters of the I/O loadand response time are analyzed in this paper. In order to achieve the purpose of forecastingresponse time by some characteristic parameter of I/O load, the model is built by combiningwith the Grey System Theory and BP Artificial Neural Network method, and the nonlinearrelations of characteristic parameter and response time is determined. Make use of the disksim,the trace data is obtained, then the trace data is used in simulating. The predicted result iscompared with the results of using grey and neural network prediction, and the advantages arehighlighted. Due to the considered factors are different, a new prediction model is proposedby combining with the BP neural network and Markov Chain, named as BP neural network-Markov chain(BP-MC) model and applied to I/O load prediction. Through emulatingthe training sample, the rolling prediction of load time series is achieved by BP neuralnetwork, and the relative error of measured and predicted values is acquired. Applying theMarkov Chain to correct the relative error and the accruacy of predicting results is improvedeffectively. The prediction model was used to predict the I/O load of Storage system， theresult shows that the model is high-precision， so it provides a new approach for Storagesystem prediction.
The evaluation of I/O performance is very important in the design and use of storagesystem. How to evaluate the I/O performance quickly and efficiently has become an importantprocedure for optimizing I/O performance. In this thesis, both qualitative and quantitativemethod is proposed, which is the combination of analytical hierarchy process level analysismethod and artificial neural network method. I/O performance is analyzed and studied by thismethod. Similarly, the trace data used for simulation is obtained by making use of the disksim.The advantage is found from the simulate result. The evaluation methodology can effectivelyevaluate the I/O performance of storage system.
引文
[1]刘凯,刘博.存储技术基础.西安:西安电子科技大学出版社,2011:1-10.
[2]刘朝斌,吴非.基于随机Petri网的虚拟存储网络性能模型与分析.第五届中国Rough集与软计算机学术研讨会.2005,32(8):285-288.
[3]向东. iSCSI-SAN网络异构存储系统管理策略的研究.华中科技大学博士学位论文.2004：1-2.
[4]谢林昂.大容量网络存储服务系统的设计与实现.上海:上海交通大学博士学位论文.2007.7-8.
[5] Theresa M. Roeder, Lee W. Schruben. Information models for queueing systemsimulation. Transactions on Modeling and Computer Simulation (TOMACS),2010,20(2):1-33.
[6] Ilias Iliadis, Cyriel Minkenberg. Performance of a speculative transmission schemefor scheduling-latency reduction. IEEE/ACM Transactions on Networking (TON),2008,16(1):182-196.
[7] Velika I. Dragieva. Some orbit size properties in the M/M/1//N retrial queue.Proceedings of the6th International Conference on Queueing Theory and NetworkApplications,2011,20(2):44-53.
[8] Gopal Sekar, Ayyappan Govindan, Muthu Ganapathi Subramanian. Stability analysisof single server retrial queueing system with Erlang service. Proceedings of the6thInternational Conference on Queueing Theory and Network,2011.155-161.
[9] Sergey Zeltyn, Yariv N. Marmor, Avishai Mandelbaum, Boaz Carmeli, OhadGreenshpan, Yossi Mesika, Sergev Wasserkrug, Pnina Vortman, Avraham Shtub,Tirza Lauterman, Dagan Schwartz, Kobi Moskovitch, Sara Tzafrir, Fuad Basis.Simulation-based models of emergency departments:: Operational, tactical, andstrategic staffing. Transactions on Modeling and Computer Simulation (TOMACS),2011,21(4):1-25.
[10]朱翼隽,朱仁祥.基于重试不耐烦M/M/s/k+M排队的呼叫中心性能分析.江苏大学学报.2004,9(1):401-404.
[11]朱翼隽,张继国,王伟.基于可变服务率M/M/s/k+M可修排队的呼叫中心性能分析.江苏大学学报.2006,27(4):368-371.
[12] Evan A. Saltzman, John H. Drew, Lawrence M. Leemis, Shane G. Henderson.Simulating Multivariate Nonhomogeneous Poisson Processes Using Projections.Transactions on Modeling and Computer Simulation (TOMACS),2012,22(3):1-13.
[13] R. Kalyanaraman. A single server batch service finite source queue with feedback.Proceedings of the6th International Conference on Queueing Theory and NetworkApplications,2011.84-88.
[14]李玺,胡志刚,阎朝坤.排队时间感知的动态网格工作流调度.湖南大学学报（自然科学版）.2012,39(3):80-87.
[15]刘朝斌.虚拟网络存储系统关键技术研究及其性能评价.华中科技大学博士学位论文.2003:1-105.
[16] C.T. Lu, D. Feng, and F. Wang. Statistical Analysis-based Approach for StorageSystem Performance Tuning. Computer Science,2010,37(11).289-294.
[17] F.Mu, W.Xue, J.W.Shu, and W.M.Zheng. An Analytical Model for Large-ScaleStorage System with Replicated Data. Journal of Computer Research andDevelopment,2009,46(5):756-761.
[18]周延年,朱怡安.基于灰熵绝对关联分析在嵌入式计算机性能评价中的应用.计算机科学,2011,38(11):206-209.
[19]宋迎迎.数字馆藏评价指标体系研究.郑州:郑州大学硕士论文.2006:1-39.
[20] Y.Zhang, Y.Zhang, and J.L.Wang. Reliability Evaluation of the Network StorageSystem Based on Block Diagrams. Computer Science,2010,37(6):102-105.
[21]张益，霍珊珊.网络存储系统可生存性量化评估.清华大学学报（自然科学版）.2009,49(82):2119-2125.
[22]冯伟.超级计算机系统性能平衡性预先评价方法研究.解放军信息工程大学硕士论文.2009:1-70.
[23]朱立谷.网络存储综合测评技术研究.计算机工程与应用.2010,16(36):61-62.
[24] L.G.Zhu, Z.S.Yang, and H.Y.Luo. Comprehensive evaluation technology fornetworked storage system. Computer Engineering and Applications,2010,46(36):61-66.
[25] G.J.Xie, J.Liu, and G.Wang. Performance Evaluation Model for Exp-RAID SystemBased on MCQN. Journal of Computer Research and Development,2008,45(suppl):207-211.
[26]陆承涛,冯丹,王芳,王娟.一种存储系统性能模型的提取方法及模型应用.小型微型计算机系统.2010,31(12):2468-2471.
[27] L.F.Wang, X.D.Zhou, Z.Q.Liu, X.M.Wang, and W.D.Liu. An Accurate EvaluationSystem for Disk Array Cache Algorithms. Journal of Northwestern PolytechnicalUniversity,2009,27(5):721-725.
[28] Yuanyuan Zhou, Zhifeng Chen, and Kai Li. Second-Level Buffer CacheManagement. IEEE Trans. Parallel Distrib.2004,6(15):505-519.
[29] Theodore M. Wong and John Wilkes. My Cache or Yours Making Storage MoreExclusive. In Proceedings of the General Track of the annual conference onUSENIX Annual Technical Conference,2002.161-175.
[30] Gala Yadgar, Michael Factor, and Assaf Schuster. Karma: know-it-all replacementfor a multilevel cache. In Proceedings of the5th USENIX conference on File andStorage Technologies,2007.25-25.
[31] Guillermo A. Alvarez, Elizabeth Borowsky, Susie Go, Theodore H. Romer, RalphBecker-Szendy, Richard Golding, Arif Merchant, Mirjana Spasojevic, AlistairVeitch, and John Wilkes. Minerva: An Automated ResourceProvisioning Tool forLarge-Scale Storage Systems. ACM Transaction on Computer System,2001,19(4):483-518.
[32] Mustafa Uysal, Guillermo A. Alvarez, Arif Merchant. A Modular, AnalyticalThroughput Model for Modern Disk Arrays. The Ninth International Symposium onModeling, Analysis and Simulation of Computer and Telecommunication Systems(MASCOTS-2001). Cincinnati, Ohio, USA,2001,15-18.
[33] Elizabeth Varki, Arif Merchant, Jianzhang Xu, Xiaozhou Qiu. Issues and Challengesin the Performance Analysis of Real Disk Arrays. IEEE Transactions on Parallel andDistributed Systems,2004,15(6):559-574.
[34] Eric Anderson. HPL–SSP–2001–4: Simple table-based modeling of storage devices.Storage and Content Distribution Department,2001,7(1):1-4.
[35] Terence Kelly, Ira Cohen, Moises Goldszmidt, Kimberly Keeton. Inducing Modelsof Black-Box Storage Arrays. Hewlett-Packard Laboratories, Palo Alto, CA, USA,2004,7(11):1-15.
[36]陆承涛.存储系统性能管理问题的研究.武汉:华中科技大学博士学位论文.2010.59-60.
[37]沈继红.灰色系统理论预测方法研究及其在舰船运动预报中的应用.哈尔滨:哈尔滨工程大学,2002.
[38]张晓龙.基于神经网络的烧结终点预测方法及应用研究.中南大学硕士学位论文.2006.23-24.
[39]赵奇,刘开第,庞彦军.灰色补偿神经网络及其应用研究[J].微计算机信息,2005，21(8)：127-128．
[40]王丽芳，周兴社，刘志强，蒋泽军，张爱华.存储系统访问特征提取系统.西北工业大学.2010,28(5):700-702.
[41]吴丽彬.面向高清媒体的嵌入式存储系统MSFS关键技术的研究.北京:中国科学院声学研究所博士学位论文.2009.
[42]朱沿旭.面向应用的缓冲区管理机制的研究与实现.湖南:国防科学技术大学硕士学位论文.2006.45-46.
[43]李兆虎，李战怀，张晓.网络存储系统仿真研究综述.计算机研究与发展.2012,49(Suppl):341-345.
[44]胡运权等著.运筹学教程.北京:清华大学出版社,1998:291-295.
[45]朱仁祥.基于M/M/S/K排队的呼叫中心.江苏大学硕士学位论文.2005:24-25.
[46]王玲.两个不同服务台的可修排队系统的矩阵几何解.燕山大学硕士学位论文.2009:5-8.
[47]余君.两个不同服务员可能故障或休假的排队系统.燕山大学理学硕士学位论文.2008,12:2-4.
[48]曹晋华,程侃等编著.可靠性数学引论.北京:科学出版社,1986:1-226.
[49]蒲宝山.较薄煤层高效开采工作面设备优化配套研究.煤炭科学研究总院博士学位论文.2006.
[50]陈洋.具有多种状态的可修M/G/1排队系统.江苏：江苏大学硕士学位论文.2003.2-3.
[51]何选森编著.随机过程.北京:人民邮电出版社,2009:34-50.
[52]殷保群,奚宏生,周亚平著.排队系统性能分析与Markov控制过程.合肥:中国科学技术大学出版社,2004:7-13.
[53]陈士成编著.运筹学：数据、模型与决策.兰州:兰州大学出版社,2009:244-269.
[54]盛友招编著.排队论及其在计算机通信中的应用.北京:北京邮电大学出版社,1998:38-41.
[55]林闯著.计算机网络和计算机系统的性能评价.北京:清华大学出版社,2001:41-45.
[56]范玉妹,徐尔,谢铁军编著.运筹学通论.北京:冶金工业出版社,2009:103-142.
[57]胡运权等著.运筹学基础及应用.第三版.哈尔滨:哈尔滨工业大学出版社,1998:193-213.
[58]唐应辉,唐小我著.排队论——基础与应用.北京:电子科技出版社,2000:35-59.
[59] Jose Blanchet, Jing Dong. Rare-event simulation for multi-server queues in theHalfin-Whitt regime. SIGMETRICS Performance Evaluation Review.2012,39(4):35-37.
[60] Jerim Kim, Jeongsim Kim, Bara Kim. Tail behavior of the queue size distribution inthe M/M/m retrial queue. Proceedings of the4th International Conference onQueueing Theory and Network Applications.2009.1-6.
[61]田志斌,田乃硕,金顺福.基于休假排队的消息驱动组件性能分析.北京工业大学学报.2011,37(10):1585-1591.
[62] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[63] Gopal Sekar, Ayyappan Govindan, Muthu Ganapathi Subramanian. Stability analysisof single server retrial queueing system with Erlang service. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011155-161.
[64] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[65] Hoda Parvin, Abhijit Bose, Mark P. Van Oyen. Priority-based routing with strictdeadlines and server flexibility under uncertainty. Winter Simulation Conference.2009.3181-3189.
[66] Joonho Kong, Sung Woo Chung, Kevin Skadron. Recent thermal managementtechniques for microprocessors. Computing Surveys (CSUR).2012,44(3):131-173.
[67] Diao Ying, Yao Nianmin. Analysis models and method on the network storagesystem I/O performance. Journal of Harbin Institute of Technology (New Series).2012,19(4):37-43.
[68] Artalejo J.R, Economou A. Markovian controllable queueing systems with hystereticpolicies: Busy period and waiting time analysis. Methodology and Computing inApplied Probability.2005,7(3):353-378.
[69] Ying DIAO, Nianmin YAO. A Kind of Modeling Based on the Storage System I/OPerformance.2012International Conference on Computer and Information Science,Safety Engineering,2012.10-12.
[70]田乃硕,岳德权著.拟生灭过程与矩阵几何解.北京:科学出版社,2002:16-26.
[71] Nail Akar, Khosrow Sohraby. Solving the single server semi-Markov queue withmatrix exponential kernel matrices for interarrivals and services. Proceedings of the1st international conference on Performance evaluation methodolgies and tools.2006.1-10.
[72] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[73]李庆海.运用拟生灭过程解决一类呼叫中心问题.首都师范大学硕士学位论文.2007:2-3,11-14.
[74] Muthu Ganapathi Subramanian, Ayyappan Govindan, Gopal Sekar. Study of multiserver retrial queueing system under vacation policies by direct truncation method.Queueing Theory and Network Applications.2011.169-177.
[75] Yong Ge, Hui Xiong, Wenjun Zhou, Siming Li, Ramendra Sahoo. Multifocallearning for customer problem analysis. Transactions on Intelligent Systems andTechnology (TIST).2011,2(3):24-46.
[76]朱翼隽.呼叫中心排队模型的研究.镇江高专学报.2006,19(4):53-58.
[77] Tasha Frankie, Gordon Hughes, Ken Kreutz-Delgado. A mathematical model of thetrim command in NAND-flash SSDs. Proceedings of the50th Annual SoutheastRegional Conference.2012.1-6.
[78]田乃硕,岳德权著.拟生灭过程与矩阵几何解.北京:科学出版社,2002:16-26.
[79] D. Bini, B. Meini, S. Steffé, J. F. Pérez, B. Van Houdt. SMCSolver and Q-MAM:tools for matrix-analytic methods. SIGMETRICS Performance Evaluation Review.2012,39(4):46-47.
[80] Song Jiang, Kei Davis, and Xiaodong Zhang. Coordinated Multilevel Buffer CacheManagement with Consistent Access Locality Quantification.2007,1(1):95-108.
[81] Asit Dan and Don Towsley.1990. An approximate analysis of the LRU and FIFObuffer replacement schemes. SIGMETRICS Perform.1990,18(1):143-152.
[82] Donghee Lee, Jongmoo Choi, Jong-Hun Kim, Sam H. Noh, Sang Lyul Min, YookunCho, and Chong Sang Kim. On the existence of a spectrum of policies that subsumesthe least recently used (LRU) and least frequently used (LFU) policies. InProceedings of the1999ACM SIGMETRICS international conference onMeasurement and modeling of computer systems (SIGMETRICS '99). ACM,1999.134-143.
[83] Elizabeth J. O'Neil, Patrick E. O'Neil, and Gerhard Weikum. The LRU-K pagereplacement algorithm for database disk buffering. In Proceedings of the1993ACMSIGMOD international conference on Management of data (SIGMOD '93), PeterBuneman and Sushil Jajodia (Eds.). ACM, New York, NY, USA,1993.297-306.
[84] Theodore Johnson and Dennis Shasha.2Q: A Low Overhead High PerformanceBuffer Management Replacement Algorithm. In Proceedings of the20thInternational Conference on Very Large Data Bases (VLDB '94), Jorge B. Bocca,Matthias Jarke, and Carlo Zaniolo (Eds.). Morgan Kaufmann Publishers Inc., SanFrancisco, CA, USA,1994.439-450.
[85] Song Jiang and Xiaodong Zhang. LIRS: an efficient low inter-reference recency setreplacement policy to improve buffer cache performance. SIGMETRICS Perform.Eval. Rev.2002,30(1):31-42.
[86] Nimrod Megiddo and Dharmendra S. Modha. ARC: A Self-Tuning, Low OverheadReplacement Cache. In Proceedings of the2nd USENIX Conference on File andStorage Technologies (FAST '03). USENIX Association, Berkeley, CA, USA,2003.115-130.
[87] Gokul Soundararajan, Jin Chen, Mohamed A. Sharaf, and Cristiana Amza. Dynamicpartitioning of the cache hierarchy in shared data centers. Proc. VLDB Endow.2008,1(1):635-646.
[88] Xuhui Li, Ashraf Aboulnaga, Kenneth Salem, Aamer Sachedina, and Shaobo Gao.Second-tier cache management using write hints. In Proceedings of the4thconference on USENIX Conference on File and Storage Technologies.2005,4(4):9-9.
[89] Bong-Jun Ko, Kang-Won Lee, Khalil Amiri, and Seraphin Calo. Scalable ServiceDifferentiation in a Shared Storage Cache. In Proceedings of the23rd InternationalConference on Distributed Computing Systems (ICDCS '03). IEEE ComputerSociety, Washington, DC, USA,2003.184-187.
[90] Z. Chen, Y. Zhou, and K. Li. Eviction-based Placement for Storage Caches. InProceedings of the USENIX Annual Technical Conference (USENIX '03), SanAntonio, Texas, June,2003.269-282.
[91] G. Edward Suh, Srinivas Devadas, and Larry Rudolph. A New Memory MonitoringScheme for Memory-Aware Scheduling and Partitioning. In Proceedings of the8thInternational Symposium on High-Performance Computer Architecture (HPCA '02).IEEE Computer Society, Washington, DC, USA,2002.117-119.
[92] Yadgar G, Factor M, Li K, Schuster A. MC2: Multiple clients on a multilevel cache.In Proc. the28th International Conference on Distributed Computing Systems(ICDCS2008),2008.722-730.
[93] Matthew Wachs, Michael Abd-El-Malek, Eno Thereska, and Gregory R. Ganger.Argon: performance insulation for shared storage servers. In Proceedings of the5thUSENIX conference on File and Storage Technologies (FAST '07). USENIXAssociation, Berkeley, CA, USA.2007.5-5.
[94] Yin Wang and Arif Merchant. Proportional-share scheduling for distributed storagesystems. In Proceedings of the5th USENIX conference on File and StorageTechnologies (FAST '07). USENIX Association, Berkeley, CA, USA,2007.4-4.
[95] Ramya Prabhakar, Shekhar Srikantaiah, Mahmut Kandemir, and Christina Patrick.Adaptive multi-level cache allocation in distributed storage architectures. InProceedings of the24th ACM International Conference on Supercomputing (ICS'10). ACM, New York, NY, USA,2010.211-221.
[96] R. H. Patterson, G. A. Gibson, E. Ginting, D. Stodolsky, and J. Zelenka. Informedprefetching and caching. SIGOPS Oper. Syst.1995,29(5):79-95.
[97] Dominique Thiebaut, Harold S. Stone, and Joel L. Wolf.1992. Improving DiskCache Hit-Ratios Through Cache Partitioning. IEEE Trans. Comput.1992,441(6):665-676.
[98] Raymond Ng, Christos Faloutsos, and Timos Sellis. Flexible and Adaptable BufferManagement Techniques for Database Management Systems. IEEE Trans. Comput.1995,44(4):546-560.
[99] J. Choi, S. Cho, S. Noh, S. Lyul, and Y. Cho, Analytical Prediction of Buffer HitRatios, Electronics Letters.2000,36(1):10-11.
[100] Jong Min Kim, Jongmoo Choi, Jesung Kim, Sam H. Noh, Sang Lyul Min, YookunCho, and Chong Sang Kim. A low-overhead high-performance unified buffermanagement scheme that exploits sequential and looping references. In Proceedingsof the4th conference on Symposium on Operating System Design\&Implementation-Volume4(OSDI'00), Vol.4. USENIX Association, Berkeley, CA,USA,2000.9-9.
[101] Thereska E, Abd-El-MalekM, Wylie J J, eta.l Informed data distribution selection ina self-predicting storage system. Proceedings of the International Conference onAutonomic Computing,2006,187-198.
[102] G. Tan, S. Wu, H. Jin, F. Xian. A scalable parallel video server based on autonomousnetwork-attached storage. Advances in Parallel Computing.2004,13(1):423-430.
[103] Xiao Qin，Performance comparisons of load balancing algorithms for I/O-intensiveworkloads on clusters Original Research Article. Journal of Network and ComputerApplications.2008,31(1):32-46.
[104] Bin Dong, Xiuqiao Li, Qimeng Wu, Limin Xiao, Li Ruan. A dynamic and adaptiveload balancing strategy for parallel file system with large-scale I/O servers. Journalof Parallel and Distributed Computing.2012,72(10):1254-1268.
[105]朱瑞祥,黄玉祥,杨晓辉.用灰色神经网络组合模型预测农机总动力发展.农业工程学报,2006,22(2):107-110.
[106]李伟,牛东晓.基于灰色神经网络的短期电力负荷预测分析.科技和产业.2008,8(10):57-59.
[107]邓聚龙编著．灰色控制系统．武汉：华中理工大学，1997：8-10.
[108]阮萍,骆力明,王华．基于灰色系统和人工神经网络的中长期电力负荷预测．首都师范大学学报（自然科学版）,2004,25(2):22-25.
[109]邓聚龙.灰色系统社会·经济.湖南:国防工业出版社,1985.
[110]彭英.基于灰色理论的数据挖掘在股票分析中的应用.湖南:长沙理工大学硕士学位论文.2006.28-29.
[111]张大海,毕研秋,毕研霞,毕研梅,牛兆水,洛鲁宁.基于串联灰色神经网络的电力负荷预测方法.系统工程理论与实践,2004,12(1):128-132.
[112]刘丹华. GM(1,1)模型的优化及应用研究.南京:南京航空航天大学硕士论文.2010.10-12.
[113]吴春广. GM(1,1)模型的改进与应用及其MATLAB实现.上海:华东师范大学硕士论文.2010.9-11.
[114]蒋宗礼.神经网络导论.北京:高等教育出版社,2001:2-20.
[115]韩仲年.基于ARM7的自适应单字体多字号识别.哈尔滨:哈尔滨工程大学硕士学位论文.2008.28-29.
[116]王冬光.控制技术在投资预测模型建立中的应用研究.哈尔滨:哈尔滨工程大学博士学位论文.2005.57-58.
[117]陈淑燕,王炜.交通量的灰色神经网络预测方法.东南大学学报(自然科学版).2004,34(4):541-544.
[118] Li Bin, Xu Shirong, Bo Guangming. Use grey-neural network combined model toforecast waste water in city. China Water&Waste Water.2002,18(2):66-68.
[119] HU Shou-ren， YU Shao-bo， DAI Kui． An introduction to neural networks．Changsha：National University of Defense Technology Press,1993:3-10.
[120]从爽．面向MATLAB工具箱的神经网络理论与应用．合肥：中国科学技术大学出版社,2009:5-48.
[121]袁曾任.人工神经网络及其应用.北京:清华大学出版社,1999:2-60.
[122]魏海坤.神经网络结构设计的理论与方法.北京:国防工业出版社,2005:7-50.
[123]杨淑娥.基于BP神经网络的上市公司财务预警模型.系统工程理论与实践.2005,1(1):14-15.
[124]吴微.神经网络计算.北京:高等教育出版社,2004:2-188.
[125]刘延锋,靳孟贵,曹英兰. BP神经网络在焉耆盆地农田排水量估算中的应用.2006,1(1):4-6.
[126]王会,清王婷,谷志红.基于灰色神经网络法的高峰负荷预测.华东电力.2005,33(4):12-13.
[127]徐卫亚.基于支持向量机-马尔可夫链的预测.科学在线.2010,31(3):944-948.
[128]曾维理,李洁,谭湘花.加权马尔可夫链在市场预测中的应用.科技资讯.2007,1(25):15-19.
[129]张宗国.马尔可夫链预测方法及其应用研究.江苏：河海大学硕士论文.2005.3-8.
[130] Robert Hulst. On the dynamics of vegetation: Markov chains as models ofsuccession,2009:169-178.
[131]阎平凡.人工神经网络与模拟进化计算.北京:清华大学出版社,2000:5-20.
[132]崔锦龙,邓姝杰.基于马尔可夫模型的降水预测及其利用.资源开发与市场.2008,24(2):115-117.
[133]张鑫.城市基础设施项目绩效评价研究.西安:西安工业大学硕士论文.2012.9-10.
[134]冯爱文.两种绿色产品评价方法在机电产品中的应用和研究.西安:西安电子科技大学硕士论文.2011.1-2.
[135]董燕.基于学习视角的秦皇岛市科技竞争力研究.秦皇岛:燕山大学硕士论文.2010.6-7.
[136]王惠荣.基于模糊综合评价理论的网上评教系统的研究与设计.华南理工大学硕士学位论文.2006.23-26.
[137]刘文娟.基于360°的煤矿安全文化模糊综合评价研究.安徽:安徽理工大学硕士论文.2011.3-4.
[138]栾进.医院医疗质量风险预警体系的构建.重庆:第三军医大学硕士论文.2011.4-5.
[139]徐茂华.我国政府以人为本评价指标研究.重庆:西南大学博士论文.2011.5-7.
[140]杨学利.基于可持续发展视角的中国粮食安全评价研究.吉林:吉林大学博士论文.2010.3-6.
[141]李友巍.网络舆论风险评估体系研究.武汉:华中师范大学硕士论文.2011.3-5.
[142]王琼.高校决策支持系统中发展评估模型研究.哈尔滨：黑龙江大学硕士论文.2005.1-4.
[143]宋培义.电视媒体数字资产管理模式研究.北京:北京交通大学博士论文.2007.120-121.
[144]孙金伟.普通高等学校教师综合评价系统的研究.大连:大连理工大学硕士学位论文.2004.24-26.
[145]张鳌.基于ANP的科技规划决策支持系统研究.北京:清华大学硕士论文.2008.2-8.
[146]沈蓉蓉.灰色理论在硬盘存储行业外包决策中的应用.上海交通大学硕士论文.2009.3-5.
[147]王文兴.虚拟计算环境中任务调度策略研究.青岛:中国石油大学.2011,2-14.
[2]刘朝斌,吴非.基于随机Petri网的虚拟存储网络性能模型与分析.第五届中国Rough集与软计算机学术研讨会.2005,32(8):285-288.
[3]向东. iSCSI-SAN网络异构存储系统管理策略的研究.华中科技大学博士学位论文.2004：1-2.
[4]谢林昂.大容量网络存储服务系统的设计与实现.上海:上海交通大学博士学位论文.2007.7-8.
[5] Theresa M. Roeder, Lee W. Schruben. Information models for queueing systemsimulation. Transactions on Modeling and Computer Simulation (TOMACS),2010,20(2):1-33.
[6] Ilias Iliadis, Cyriel Minkenberg. Performance of a speculative transmission schemefor scheduling-latency reduction. IEEE/ACM Transactions on Networking (TON),2008,16(1):182-196.
[7] Velika I. Dragieva. Some orbit size properties in the M/M/1//N retrial queue.Proceedings of the6th International Conference on Queueing Theory and NetworkApplications,2011,20(2):44-53.
[8] Gopal Sekar, Ayyappan Govindan, Muthu Ganapathi Subramanian. Stability analysisof single server retrial queueing system with Erlang service. Proceedings of the6thInternational Conference on Queueing Theory and Network,2011.155-161.
[9] Sergey Zeltyn, Yariv N. Marmor, Avishai Mandelbaum, Boaz Carmeli, OhadGreenshpan, Yossi Mesika, Sergev Wasserkrug, Pnina Vortman, Avraham Shtub,Tirza Lauterman, Dagan Schwartz, Kobi Moskovitch, Sara Tzafrir, Fuad Basis.Simulation-based models of emergency departments:: Operational, tactical, andstrategic staffing. Transactions on Modeling and Computer Simulation (TOMACS),2011,21(4):1-25.
[10]朱翼隽,朱仁祥.基于重试不耐烦M/M/s/k+M排队的呼叫中心性能分析.江苏大学学报.2004,9(1):401-404.
[11]朱翼隽,张继国,王伟.基于可变服务率M/M/s/k+M可修排队的呼叫中心性能分析.江苏大学学报.2006,27(4):368-371.
[12] Evan A. Saltzman, John H. Drew, Lawrence M. Leemis, Shane G. Henderson.Simulating Multivariate Nonhomogeneous Poisson Processes Using Projections.Transactions on Modeling and Computer Simulation (TOMACS),2012,22(3):1-13.
[13] R. Kalyanaraman. A single server batch service finite source queue with feedback.Proceedings of the6th International Conference on Queueing Theory and NetworkApplications,2011.84-88.
[14]李玺,胡志刚,阎朝坤.排队时间感知的动态网格工作流调度.湖南大学学报（自然科学版）.2012,39(3):80-87.
[15]刘朝斌.虚拟网络存储系统关键技术研究及其性能评价.华中科技大学博士学位论文.2003:1-105.
[16] C.T. Lu, D. Feng, and F. Wang. Statistical Analysis-based Approach for StorageSystem Performance Tuning. Computer Science,2010,37(11).289-294.
[17] F.Mu, W.Xue, J.W.Shu, and W.M.Zheng. An Analytical Model for Large-ScaleStorage System with Replicated Data. Journal of Computer Research andDevelopment,2009,46(5):756-761.
[18]周延年,朱怡安.基于灰熵绝对关联分析在嵌入式计算机性能评价中的应用.计算机科学,2011,38(11):206-209.
[19]宋迎迎.数字馆藏评价指标体系研究.郑州:郑州大学硕士论文.2006:1-39.
[20] Y.Zhang, Y.Zhang, and J.L.Wang. Reliability Evaluation of the Network StorageSystem Based on Block Diagrams. Computer Science,2010,37(6):102-105.
[21]张益，霍珊珊.网络存储系统可生存性量化评估.清华大学学报（自然科学版）.2009,49(82):2119-2125.
[22]冯伟.超级计算机系统性能平衡性预先评价方法研究.解放军信息工程大学硕士论文.2009:1-70.
[23]朱立谷.网络存储综合测评技术研究.计算机工程与应用.2010,16(36):61-62.
[24] L.G.Zhu, Z.S.Yang, and H.Y.Luo. Comprehensive evaluation technology fornetworked storage system. Computer Engineering and Applications,2010,46(36):61-66.
[25] G.J.Xie, J.Liu, and G.Wang. Performance Evaluation Model for Exp-RAID SystemBased on MCQN. Journal of Computer Research and Development,2008,45(suppl):207-211.
[26]陆承涛,冯丹,王芳,王娟.一种存储系统性能模型的提取方法及模型应用.小型微型计算机系统.2010,31(12):2468-2471.
[27] L.F.Wang, X.D.Zhou, Z.Q.Liu, X.M.Wang, and W.D.Liu. An Accurate EvaluationSystem for Disk Array Cache Algorithms. Journal of Northwestern PolytechnicalUniversity,2009,27(5):721-725.
[28] Yuanyuan Zhou, Zhifeng Chen, and Kai Li. Second-Level Buffer CacheManagement. IEEE Trans. Parallel Distrib.2004,6(15):505-519.
[29] Theodore M. Wong and John Wilkes. My Cache or Yours Making Storage MoreExclusive. In Proceedings of the General Track of the annual conference onUSENIX Annual Technical Conference,2002.161-175.
[30] Gala Yadgar, Michael Factor, and Assaf Schuster. Karma: know-it-all replacementfor a multilevel cache. In Proceedings of the5th USENIX conference on File andStorage Technologies,2007.25-25.
[31] Guillermo A. Alvarez, Elizabeth Borowsky, Susie Go, Theodore H. Romer, RalphBecker-Szendy, Richard Golding, Arif Merchant, Mirjana Spasojevic, AlistairVeitch, and John Wilkes. Minerva: An Automated ResourceProvisioning Tool forLarge-Scale Storage Systems. ACM Transaction on Computer System,2001,19(4):483-518.
[32] Mustafa Uysal, Guillermo A. Alvarez, Arif Merchant. A Modular, AnalyticalThroughput Model for Modern Disk Arrays. The Ninth International Symposium onModeling, Analysis and Simulation of Computer and Telecommunication Systems(MASCOTS-2001). Cincinnati, Ohio, USA,2001,15-18.
[33] Elizabeth Varki, Arif Merchant, Jianzhang Xu, Xiaozhou Qiu. Issues and Challengesin the Performance Analysis of Real Disk Arrays. IEEE Transactions on Parallel andDistributed Systems,2004,15(6):559-574.
[34] Eric Anderson. HPL–SSP–2001–4: Simple table-based modeling of storage devices.Storage and Content Distribution Department,2001,7(1):1-4.
[35] Terence Kelly, Ira Cohen, Moises Goldszmidt, Kimberly Keeton. Inducing Modelsof Black-Box Storage Arrays. Hewlett-Packard Laboratories, Palo Alto, CA, USA,2004,7(11):1-15.
[36]陆承涛.存储系统性能管理问题的研究.武汉:华中科技大学博士学位论文.2010.59-60.
[37]沈继红.灰色系统理论预测方法研究及其在舰船运动预报中的应用.哈尔滨:哈尔滨工程大学,2002.
[38]张晓龙.基于神经网络的烧结终点预测方法及应用研究.中南大学硕士学位论文.2006.23-24.
[39]赵奇,刘开第,庞彦军.灰色补偿神经网络及其应用研究[J].微计算机信息,2005，21(8)：127-128．
[40]王丽芳，周兴社，刘志强，蒋泽军，张爱华.存储系统访问特征提取系统.西北工业大学.2010,28(5):700-702.
[41]吴丽彬.面向高清媒体的嵌入式存储系统MSFS关键技术的研究.北京:中国科学院声学研究所博士学位论文.2009.
[42]朱沿旭.面向应用的缓冲区管理机制的研究与实现.湖南:国防科学技术大学硕士学位论文.2006.45-46.
[43]李兆虎，李战怀，张晓.网络存储系统仿真研究综述.计算机研究与发展.2012,49(Suppl):341-345.
[44]胡运权等著.运筹学教程.北京:清华大学出版社,1998:291-295.
[45]朱仁祥.基于M/M/S/K排队的呼叫中心.江苏大学硕士学位论文.2005:24-25.
[46]王玲.两个不同服务台的可修排队系统的矩阵几何解.燕山大学硕士学位论文.2009:5-8.
[47]余君.两个不同服务员可能故障或休假的排队系统.燕山大学理学硕士学位论文.2008,12:2-4.
[48]曹晋华,程侃等编著.可靠性数学引论.北京:科学出版社,1986:1-226.
[49]蒲宝山.较薄煤层高效开采工作面设备优化配套研究.煤炭科学研究总院博士学位论文.2006.
[50]陈洋.具有多种状态的可修M/G/1排队系统.江苏：江苏大学硕士学位论文.2003.2-3.
[51]何选森编著.随机过程.北京:人民邮电出版社,2009:34-50.
[52]殷保群,奚宏生,周亚平著.排队系统性能分析与Markov控制过程.合肥:中国科学技术大学出版社,2004:7-13.
[53]陈士成编著.运筹学：数据、模型与决策.兰州:兰州大学出版社,2009:244-269.
[54]盛友招编著.排队论及其在计算机通信中的应用.北京:北京邮电大学出版社,1998:38-41.
[55]林闯著.计算机网络和计算机系统的性能评价.北京:清华大学出版社,2001:41-45.
[56]范玉妹,徐尔,谢铁军编著.运筹学通论.北京:冶金工业出版社,2009:103-142.
[57]胡运权等著.运筹学基础及应用.第三版.哈尔滨:哈尔滨工业大学出版社,1998:193-213.
[58]唐应辉,唐小我著.排队论——基础与应用.北京:电子科技出版社,2000:35-59.
[59] Jose Blanchet, Jing Dong. Rare-event simulation for multi-server queues in theHalfin-Whitt regime. SIGMETRICS Performance Evaluation Review.2012,39(4):35-37.
[60] Jerim Kim, Jeongsim Kim, Bara Kim. Tail behavior of the queue size distribution inthe M/M/m retrial queue. Proceedings of the4th International Conference onQueueing Theory and Network Applications.2009.1-6.
[61]田志斌,田乃硕,金顺福.基于休假排队的消息驱动组件性能分析.北京工业大学学报.2011,37(10):1585-1591.
[62] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[63] Gopal Sekar, Ayyappan Govindan, Muthu Ganapathi Subramanian. Stability analysisof single server retrial queueing system with Erlang service. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011155-161.
[64] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[65] Hoda Parvin, Abhijit Bose, Mark P. Van Oyen. Priority-based routing with strictdeadlines and server flexibility under uncertainty. Winter Simulation Conference.2009.3181-3189.
[66] Joonho Kong, Sung Woo Chung, Kevin Skadron. Recent thermal managementtechniques for microprocessors. Computing Surveys (CSUR).2012,44(3):131-173.
[67] Diao Ying, Yao Nianmin. Analysis models and method on the network storagesystem I/O performance. Journal of Harbin Institute of Technology (New Series).2012,19(4):37-43.
[68] Artalejo J.R, Economou A. Markovian controllable queueing systems with hystereticpolicies: Busy period and waiting time analysis. Methodology and Computing inApplied Probability.2005,7(3):353-378.
[69] Ying DIAO, Nianmin YAO. A Kind of Modeling Based on the Storage System I/OPerformance.2012International Conference on Computer and Information Science,Safety Engineering,2012.10-12.
[70]田乃硕,岳德权著.拟生灭过程与矩阵几何解.北京:科学出版社,2002:16-26.
[71] Nail Akar, Khosrow Sohraby. Solving the single server semi-Markov queue withmatrix exponential kernel matrices for interarrivals and services. Proceedings of the1st international conference on Performance evaluation methodolgies and tools.2006.1-10.
[72] Kaliappan Kalidass, Kasturi Ramanath. A priority retrial queue with second multioptional service and m immediate Bernoulli feedbacks. Proceedings of the6thInternational Conference on Queueing Theory and Network Applications.2011.67-77.
[73]李庆海.运用拟生灭过程解决一类呼叫中心问题.首都师范大学硕士学位论文.2007:2-3,11-14.
[74] Muthu Ganapathi Subramanian, Ayyappan Govindan, Gopal Sekar. Study of multiserver retrial queueing system under vacation policies by direct truncation method.Queueing Theory and Network Applications.2011.169-177.
[75] Yong Ge, Hui Xiong, Wenjun Zhou, Siming Li, Ramendra Sahoo. Multifocallearning for customer problem analysis. Transactions on Intelligent Systems andTechnology (TIST).2011,2(3):24-46.
[76]朱翼隽.呼叫中心排队模型的研究.镇江高专学报.2006,19(4):53-58.
[77] Tasha Frankie, Gordon Hughes, Ken Kreutz-Delgado. A mathematical model of thetrim command in NAND-flash SSDs. Proceedings of the50th Annual SoutheastRegional Conference.2012.1-6.
[78]田乃硕,岳德权著.拟生灭过程与矩阵几何解.北京:科学出版社,2002:16-26.
[79] D. Bini, B. Meini, S. Steffé, J. F. Pérez, B. Van Houdt. SMCSolver and Q-MAM:tools for matrix-analytic methods. SIGMETRICS Performance Evaluation Review.2012,39(4):46-47.
[80] Song Jiang, Kei Davis, and Xiaodong Zhang. Coordinated Multilevel Buffer CacheManagement with Consistent Access Locality Quantification.2007,1(1):95-108.
[81] Asit Dan and Don Towsley.1990. An approximate analysis of the LRU and FIFObuffer replacement schemes. SIGMETRICS Perform.1990,18(1):143-152.
[82] Donghee Lee, Jongmoo Choi, Jong-Hun Kim, Sam H. Noh, Sang Lyul Min, YookunCho, and Chong Sang Kim. On the existence of a spectrum of policies that subsumesthe least recently used (LRU) and least frequently used (LFU) policies. InProceedings of the1999ACM SIGMETRICS international conference onMeasurement and modeling of computer systems (SIGMETRICS '99). ACM,1999.134-143.
[83] Elizabeth J. O'Neil, Patrick E. O'Neil, and Gerhard Weikum. The LRU-K pagereplacement algorithm for database disk buffering. In Proceedings of the1993ACMSIGMOD international conference on Management of data (SIGMOD '93), PeterBuneman and Sushil Jajodia (Eds.). ACM, New York, NY, USA,1993.297-306.
[84] Theodore Johnson and Dennis Shasha.2Q: A Low Overhead High PerformanceBuffer Management Replacement Algorithm. In Proceedings of the20thInternational Conference on Very Large Data Bases (VLDB '94), Jorge B. Bocca,Matthias Jarke, and Carlo Zaniolo (Eds.). Morgan Kaufmann Publishers Inc., SanFrancisco, CA, USA,1994.439-450.
[85] Song Jiang and Xiaodong Zhang. LIRS: an efficient low inter-reference recency setreplacement policy to improve buffer cache performance. SIGMETRICS Perform.Eval. Rev.2002,30(1):31-42.
[86] Nimrod Megiddo and Dharmendra S. Modha. ARC: A Self-Tuning, Low OverheadReplacement Cache. In Proceedings of the2nd USENIX Conference on File andStorage Technologies (FAST '03). USENIX Association, Berkeley, CA, USA,2003.115-130.
[87] Gokul Soundararajan, Jin Chen, Mohamed A. Sharaf, and Cristiana Amza. Dynamicpartitioning of the cache hierarchy in shared data centers. Proc. VLDB Endow.2008,1(1):635-646.
[88] Xuhui Li, Ashraf Aboulnaga, Kenneth Salem, Aamer Sachedina, and Shaobo Gao.Second-tier cache management using write hints. In Proceedings of the4thconference on USENIX Conference on File and Storage Technologies.2005,4(4):9-9.
[89] Bong-Jun Ko, Kang-Won Lee, Khalil Amiri, and Seraphin Calo. Scalable ServiceDifferentiation in a Shared Storage Cache. In Proceedings of the23rd InternationalConference on Distributed Computing Systems (ICDCS '03). IEEE ComputerSociety, Washington, DC, USA,2003.184-187.
[90] Z. Chen, Y. Zhou, and K. Li. Eviction-based Placement for Storage Caches. InProceedings of the USENIX Annual Technical Conference (USENIX '03), SanAntonio, Texas, June,2003.269-282.
[91] G. Edward Suh, Srinivas Devadas, and Larry Rudolph. A New Memory MonitoringScheme for Memory-Aware Scheduling and Partitioning. In Proceedings of the8thInternational Symposium on High-Performance Computer Architecture (HPCA '02).IEEE Computer Society, Washington, DC, USA,2002.117-119.
[92] Yadgar G, Factor M, Li K, Schuster A. MC2: Multiple clients on a multilevel cache.In Proc. the28th International Conference on Distributed Computing Systems(ICDCS2008),2008.722-730.
[93] Matthew Wachs, Michael Abd-El-Malek, Eno Thereska, and Gregory R. Ganger.Argon: performance insulation for shared storage servers. In Proceedings of the5thUSENIX conference on File and Storage Technologies (FAST '07). USENIXAssociation, Berkeley, CA, USA.2007.5-5.
[94] Yin Wang and Arif Merchant. Proportional-share scheduling for distributed storagesystems. In Proceedings of the5th USENIX conference on File and StorageTechnologies (FAST '07). USENIX Association, Berkeley, CA, USA,2007.4-4.
[95] Ramya Prabhakar, Shekhar Srikantaiah, Mahmut Kandemir, and Christina Patrick.Adaptive multi-level cache allocation in distributed storage architectures. InProceedings of the24th ACM International Conference on Supercomputing (ICS'10). ACM, New York, NY, USA,2010.211-221.
[96] R. H. Patterson, G. A. Gibson, E. Ginting, D. Stodolsky, and J. Zelenka. Informedprefetching and caching. SIGOPS Oper. Syst.1995,29(5):79-95.
[97] Dominique Thiebaut, Harold S. Stone, and Joel L. Wolf.1992. Improving DiskCache Hit-Ratios Through Cache Partitioning. IEEE Trans. Comput.1992,441(6):665-676.
[98] Raymond Ng, Christos Faloutsos, and Timos Sellis. Flexible and Adaptable BufferManagement Techniques for Database Management Systems. IEEE Trans. Comput.1995,44(4):546-560.
[99] J. Choi, S. Cho, S. Noh, S. Lyul, and Y. Cho, Analytical Prediction of Buffer HitRatios, Electronics Letters.2000,36(1):10-11.
[100] Jong Min Kim, Jongmoo Choi, Jesung Kim, Sam H. Noh, Sang Lyul Min, YookunCho, and Chong Sang Kim. A low-overhead high-performance unified buffermanagement scheme that exploits sequential and looping references. In Proceedingsof the4th conference on Symposium on Operating System Design\&Implementation-Volume4(OSDI'00), Vol.4. USENIX Association, Berkeley, CA,USA,2000.9-9.
[101] Thereska E, Abd-El-MalekM, Wylie J J, eta.l Informed data distribution selection ina self-predicting storage system. Proceedings of the International Conference onAutonomic Computing,2006,187-198.
[102] G. Tan, S. Wu, H. Jin, F. Xian. A scalable parallel video server based on autonomousnetwork-attached storage. Advances in Parallel Computing.2004,13(1):423-430.
[103] Xiao Qin，Performance comparisons of load balancing algorithms for I/O-intensiveworkloads on clusters Original Research Article. Journal of Network and ComputerApplications.2008,31(1):32-46.
[104] Bin Dong, Xiuqiao Li, Qimeng Wu, Limin Xiao, Li Ruan. A dynamic and adaptiveload balancing strategy for parallel file system with large-scale I/O servers. Journalof Parallel and Distributed Computing.2012,72(10):1254-1268.
[105]朱瑞祥,黄玉祥,杨晓辉.用灰色神经网络组合模型预测农机总动力发展.农业工程学报,2006,22(2):107-110.
[106]李伟,牛东晓.基于灰色神经网络的短期电力负荷预测分析.科技和产业.2008,8(10):57-59.
[107]邓聚龙编著．灰色控制系统．武汉：华中理工大学，1997：8-10.
[108]阮萍,骆力明,王华．基于灰色系统和人工神经网络的中长期电力负荷预测．首都师范大学学报（自然科学版）,2004,25(2):22-25.
[109]邓聚龙.灰色系统社会·经济.湖南:国防工业出版社,1985.
[110]彭英.基于灰色理论的数据挖掘在股票分析中的应用.湖南:长沙理工大学硕士学位论文.2006.28-29.
[111]张大海,毕研秋,毕研霞,毕研梅,牛兆水,洛鲁宁.基于串联灰色神经网络的电力负荷预测方法.系统工程理论与实践,2004,12(1):128-132.
[112]刘丹华. GM(1,1)模型的优化及应用研究.南京:南京航空航天大学硕士论文.2010.10-12.
[113]吴春广. GM(1,1)模型的改进与应用及其MATLAB实现.上海:华东师范大学硕士论文.2010.9-11.
[114]蒋宗礼.神经网络导论.北京:高等教育出版社,2001:2-20.
[115]韩仲年.基于ARM7的自适应单字体多字号识别.哈尔滨:哈尔滨工程大学硕士学位论文.2008.28-29.
[116]王冬光.控制技术在投资预测模型建立中的应用研究.哈尔滨:哈尔滨工程大学博士学位论文.2005.57-58.
[117]陈淑燕,王炜.交通量的灰色神经网络预测方法.东南大学学报(自然科学版).2004,34(4):541-544.
[118] Li Bin, Xu Shirong, Bo Guangming. Use grey-neural network combined model toforecast waste water in city. China Water&Waste Water.2002,18(2):66-68.
[119] HU Shou-ren， YU Shao-bo， DAI Kui． An introduction to neural networks．Changsha：National University of Defense Technology Press,1993:3-10.
[120]从爽．面向MATLAB工具箱的神经网络理论与应用．合肥：中国科学技术大学出版社,2009:5-48.
[121]袁曾任.人工神经网络及其应用.北京:清华大学出版社,1999:2-60.
[122]魏海坤.神经网络结构设计的理论与方法.北京:国防工业出版社,2005:7-50.
[123]杨淑娥.基于BP神经网络的上市公司财务预警模型.系统工程理论与实践.2005,1(1):14-15.
[124]吴微.神经网络计算.北京:高等教育出版社,2004:2-188.
[125]刘延锋,靳孟贵,曹英兰. BP神经网络在焉耆盆地农田排水量估算中的应用.2006,1(1):4-6.
[126]王会,清王婷,谷志红.基于灰色神经网络法的高峰负荷预测.华东电力.2005,33(4):12-13.
[127]徐卫亚.基于支持向量机-马尔可夫链的预测.科学在线.2010,31(3):944-948.
[128]曾维理,李洁,谭湘花.加权马尔可夫链在市场预测中的应用.科技资讯.2007,1(25):15-19.
[129]张宗国.马尔可夫链预测方法及其应用研究.江苏：河海大学硕士论文.2005.3-8.
[130] Robert Hulst. On the dynamics of vegetation: Markov chains as models ofsuccession,2009:169-178.
[131]阎平凡.人工神经网络与模拟进化计算.北京:清华大学出版社,2000:5-20.
[132]崔锦龙,邓姝杰.基于马尔可夫模型的降水预测及其利用.资源开发与市场.2008,24(2):115-117.
[133]张鑫.城市基础设施项目绩效评价研究.西安:西安工业大学硕士论文.2012.9-10.
[134]冯爱文.两种绿色产品评价方法在机电产品中的应用和研究.西安:西安电子科技大学硕士论文.2011.1-2.
[135]董燕.基于学习视角的秦皇岛市科技竞争力研究.秦皇岛:燕山大学硕士论文.2010.6-7.
[136]王惠荣.基于模糊综合评价理论的网上评教系统的研究与设计.华南理工大学硕士学位论文.2006.23-26.
[137]刘文娟.基于360°的煤矿安全文化模糊综合评价研究.安徽:安徽理工大学硕士论文.2011.3-4.
[138]栾进.医院医疗质量风险预警体系的构建.重庆:第三军医大学硕士论文.2011.4-5.
[139]徐茂华.我国政府以人为本评价指标研究.重庆:西南大学博士论文.2011.5-7.
[140]杨学利.基于可持续发展视角的中国粮食安全评价研究.吉林:吉林大学博士论文.2010.3-6.
[141]李友巍.网络舆论风险评估体系研究.武汉:华中师范大学硕士论文.2011.3-5.
[142]王琼.高校决策支持系统中发展评估模型研究.哈尔滨：黑龙江大学硕士论文.2005.1-4.
[143]宋培义.电视媒体数字资产管理模式研究.北京:北京交通大学博士论文.2007.120-121.
[144]孙金伟.普通高等学校教师综合评价系统的研究.大连:大连理工大学硕士学位论文.2004.24-26.
[145]张鳌.基于ANP的科技规划决策支持系统研究.北京:清华大学硕士论文.2008.2-8.
[146]沈蓉蓉.灰色理论在硬盘存储行业外包决策中的应用.上海交通大学硕士论文.2009.3-5.
[147]王文兴.虚拟计算环境中任务调度策略研究.青岛:中国石油大学.2011,2-14.