面向高性能计算的可扩展I/O体系结构研究与实现
|
摘要
数值模拟计算是进行科学研究和探索的主要技术手段之一,其对计算机的计算和数据处理能力提出了巨大的、不断增长的需求,推动着并行计算机系统的发展。高性能计算已进入PetaFlops时代,与此同时,数据存储也进入Petabyte(千万亿字节)时代,对I/O性能、可扩展性、可靠性、可用性和易管理性提出了严峻挑战。大规模并行计算机系统的I/O效能已经成为阻碍系统获得高效能的重要瓶颈。这主要表现在两个方面,一是I/O设备速度、I/O体系结构等因素的制约,使系统I/O性能和计算性能严重不匹配;二是系统规模的扩展导致I/O设备高故障率和数据恢复时间的增长,使I/O系统的可用性问题日益突出。
为了缓解I/O瓶颈问题,可以从应用程序、可扩展算法、编译器和语言、运行时库、操作系统和体系结构六方面展开研究。其中,I/O体系结构是所有技术途径的关键支撑。针对高性能计算I/O需求与挑战,结合高效能并行计算机系统的研制任务,论文首先研究了I/O体系结构,从体系结构上保证并行I/O的性能及可扩展性。其次,在实现机制上,研究了涵盖I/O的存储一致性模型及实现技术、智能I/O控制、电磁混合存储加速和事务型存储管理等技术,达到提高并行I/O性能和系统可用性的目的。
论文的主要研究工作和创新点如下:
1. I/O受限的并行加速比模型
针对并行计算机系统的可扩展性问题,研究了I/O负载对并行计算机系统可扩展性的影响,提出了I/O受限的并行加速比性能模型,以此为基础对三类常见的I/O体系结构的可扩展性进行了分析;最后用性能模型指导I/O体系结构设计,设计了一种面向高性能计算的可扩展并行I/O系统结构,提出了提高系统可扩展性的几种策略。
2.涵盖I/O的广义域存储一致性模型及协议实现技术
针对支持全局DMA操作的共享存储系统存储一致性问题,从I/O与存储体系结构一体化设计理念出发,定义了涵盖I/O的广义程序概念,研究了广义存储一致性,建立了广义顺序一致性模型、广义释放一致性模型和广义域一致性模型,基于广义域一致性模型设计并实现了Cache-Memory-I/O数据一致性协议,在大规模CC-NUMA系统上实现了支持全局并发DMA的全局共享I/O系统。实测结果表明,该系统I/O吞吐能力和扩展能力强,实测并行I/O带宽高达20.2GB/s,并行I/O带宽随着进程个数良好扩展。
3.基于强化学习的智能I/O调度算法RL-scheduler
针对实际应用中磁盘阵列的I/O服务效率问题,将机器学习领域中的强化学习技术引入RAID控制器中,提出了基于强化学习的智能I/O调度算法RL-scheduler,利用Q-学习策略实现了面向并行应用的自治调度策略。RL-scheduler综合考虑了调度的公平性、磁盘寻道时间和MPI应用的I/O访问效率,并提出多Q-表交叉组织方法提高Q-表的更新效率。实验结果表明,RL-scheduler缩短了并行应用的平均I/O等待时间,提高了大规模并行计算机系统的实用I/O带宽,增强了系统的可扩展性。
4.支持事务语义的电磁混合存储管理算法
针对高性能计算对I/O性能和可用性的双重需求与挑战,在存储设备一级将事务型存储管理和电磁混合存储加速技术有机结合,研究支持事务语义的电磁混合存储技术,提出了基于令牌的并行事务冲突处理协议和自适应动态逻辑分区管理算法。模拟结果表明,支持事务语义的电磁混合存储系统能够有效利用事务访问规律提高固态盘缓存命中率,隐藏版本管理、冲突检测等开销,获得I/O性能和可用性的双重改善。
Numerical simulation computing serve as one of the main methods to performscientificresearch and exploration, which poses a tremendous and continuous growingdemand on the computation and data processing capacity of high performancecomputers, has driven the development of scientific computing and parallel computersystem. At present, High Performance Computing has entered an era of Petaflops, andthe storage systems also entered the Petabyte era. The challenges of petascalecomputingondatastoragecapacity, I/Operformance,scalability,reliability, availability,and manageability are tremendous. However, the I/O bottleneck issues obstruct largescale parallel systems to achieve higher efficiency, which happens in two occasions. Inthefirst place, I/O performanceis restricted byfactors such as I/Odevice speedand I/Oarchitectures, which results in I/O and computing speed beingsignificantly unmatched.Secondly, scaling up system size makes disk drivefailure more frequently and longertime to reconstruct the failed drive; in consequence availabilityof I/O system becomesmuchcriticalissue.
The effective solutions for the I/O bottleneck can be found from the following sixlevels, including applications, algorithms, languages and compilers, run-time libraries,operating systems, and I/O architecture. Among all the levels mentioned above, I/Oarchitectureisthemostfundamental.
Inordertomeet the I/Orequirement andchallenge,alongwithourresearchtask ofa high performance parallel computing system, this paper is presenting our theoreticalstudy of I/O architectures, from which make it possible the high performance andscalability in terms of I/O architecture level. Meantime, I/O implementationmechanisms is focused on this paper, including technologies such as I/O-includedmemory consistency model and its implementation, intelligent I/O control, hybridstorage and transactional storage management, so as to promote I/O performance andavailability.
Themainworkandinnovativepointsofthispaperareasfollows.
1. I/Orestrictedparallelspeedup model
Current parallel I/O performance analysis lacks scientific theoretical models tosupporttheI/Oarchitecturedesign.ThepaperstudiestheimpactofI/Oworkloadonthescalability of parallel computing systems and proposes the I/O restricted parallelspeedupmodel.Basedonthismodel,whichcanbeusedtoguideI/Oarchitecturedesign,a scalable parallel I/O architecture for HPC is presented. Moreover, the paper analyzesseveral strategies for improving the system scalability, which serve as the basis forfurtherstudy.
2. I/O-includedgeneralmemoryconsistencymodelandimplementingtechnology
As for the consistency problem of shared memory systems with global DMAoperations,thepaperdefinestheconceptof I/O-includedgeneral program.Basedontheconcept, the paper studies the general memory consistency model, builds the generalsequence consistency model, general release consistency model and general scopeconsistency model. Using general scope memory consistency model, the paper designsand implements the CC-NUMA Cache Coherence protocol with global DMA and theglobal shared parallel I/O architecture at the hardware level. The experiment resultsshow that the I/O bandwidth and scalability of the system perform fairly well. Theactual parallel I/O bandwidth reaches 20.2 GB/s, and scales well with the number ofsystemprocesses.
3. IntelligentI/Oschedulealgorithmbasedonreinforcementlearning
To improve the I/O service efficiency of RAID and optimize the I/O performanceof parallel applications, the paper presents an intelligent I/O schedule algorithm,RL-scheduler, in RAID controllers based on reinforcement learning. RL-schedulerutilizes Q-learning strategy to implement a self-control and self-optimizing scheduler.The algorithm leverages the scheduling equity, disk seeking time and the I/O accessefficiency of MPI applications. Furthermore, the proposed interleaving organization ofmultiple Q-tables improves the efficiency of the Q-table updating. The experimentresults show that, on a large-scale parallel system with multiple parallel applications,RL-scheduler shortens the average I/O waiting time of parallel applicationsconsiderably. Thus increases the practical I/O bandwidth, and improves the systems’scalability.
4. Hybrid storagemanagement algorithmtosupport transactionsemantics
To address the requirement and challenge posed by HPC, the paper combines theidea of transactional storage management and hybrid storage acceleration, andintroduces an electro-magnetic hybrid storage management algorithm to supporttransaction semantics. A token-based protocol is designed to cope with the conflictsbetween I/O transactions and an adaptive logical partition algorithm is proposed tomanage Solid State Disk (SSD) storage. Simulation results show that theelectro-magnetic hybrid storage system can deal with transactions with varied accesspattern elegantly and effectively improve SSD hit rate, hide the overhead of versionmanagement and conflict detection. Both the I/O performance and availability aresignificantlyimproved.
为了缓解I/O瓶颈问题,可以从应用程序、可扩展算法、编译器和语言、运行时库、操作系统和体系结构六方面展开研究。其中,I/O体系结构是所有技术途径的关键支撑。针对高性能计算I/O需求与挑战,结合高效能并行计算机系统的研制任务,论文首先研究了I/O体系结构,从体系结构上保证并行I/O的性能及可扩展性。其次,在实现机制上,研究了涵盖I/O的存储一致性模型及实现技术、智能I/O控制、电磁混合存储加速和事务型存储管理等技术,达到提高并行I/O性能和系统可用性的目的。
论文的主要研究工作和创新点如下:
1. I/O受限的并行加速比模型
针对并行计算机系统的可扩展性问题,研究了I/O负载对并行计算机系统可扩展性的影响,提出了I/O受限的并行加速比性能模型,以此为基础对三类常见的I/O体系结构的可扩展性进行了分析;最后用性能模型指导I/O体系结构设计,设计了一种面向高性能计算的可扩展并行I/O系统结构,提出了提高系统可扩展性的几种策略。
2.涵盖I/O的广义域存储一致性模型及协议实现技术
针对支持全局DMA操作的共享存储系统存储一致性问题,从I/O与存储体系结构一体化设计理念出发,定义了涵盖I/O的广义程序概念,研究了广义存储一致性,建立了广义顺序一致性模型、广义释放一致性模型和广义域一致性模型,基于广义域一致性模型设计并实现了Cache-Memory-I/O数据一致性协议,在大规模CC-NUMA系统上实现了支持全局并发DMA的全局共享I/O系统。实测结果表明,该系统I/O吞吐能力和扩展能力强,实测并行I/O带宽高达20.2GB/s,并行I/O带宽随着进程个数良好扩展。
3.基于强化学习的智能I/O调度算法RL-scheduler
针对实际应用中磁盘阵列的I/O服务效率问题,将机器学习领域中的强化学习技术引入RAID控制器中,提出了基于强化学习的智能I/O调度算法RL-scheduler,利用Q-学习策略实现了面向并行应用的自治调度策略。RL-scheduler综合考虑了调度的公平性、磁盘寻道时间和MPI应用的I/O访问效率,并提出多Q-表交叉组织方法提高Q-表的更新效率。实验结果表明,RL-scheduler缩短了并行应用的平均I/O等待时间,提高了大规模并行计算机系统的实用I/O带宽,增强了系统的可扩展性。
4.支持事务语义的电磁混合存储管理算法
针对高性能计算对I/O性能和可用性的双重需求与挑战,在存储设备一级将事务型存储管理和电磁混合存储加速技术有机结合,研究支持事务语义的电磁混合存储技术,提出了基于令牌的并行事务冲突处理协议和自适应动态逻辑分区管理算法。模拟结果表明,支持事务语义的电磁混合存储系统能够有效利用事务访问规律提高固态盘缓存命中率,隐藏版本管理、冲突检测等开销,获得I/O性能和可用性的双重改善。
Numerical simulation computing serve as one of the main methods to performscientificresearch and exploration, which poses a tremendous and continuous growingdemand on the computation and data processing capacity of high performancecomputers, has driven the development of scientific computing and parallel computersystem. At present, High Performance Computing has entered an era of Petaflops, andthe storage systems also entered the Petabyte era. The challenges of petascalecomputingondatastoragecapacity, I/Operformance,scalability,reliability, availability,and manageability are tremendous. However, the I/O bottleneck issues obstruct largescale parallel systems to achieve higher efficiency, which happens in two occasions. Inthefirst place, I/O performanceis restricted byfactors such as I/Odevice speedand I/Oarchitectures, which results in I/O and computing speed beingsignificantly unmatched.Secondly, scaling up system size makes disk drivefailure more frequently and longertime to reconstruct the failed drive; in consequence availabilityof I/O system becomesmuchcriticalissue.
The effective solutions for the I/O bottleneck can be found from the following sixlevels, including applications, algorithms, languages and compilers, run-time libraries,operating systems, and I/O architecture. Among all the levels mentioned above, I/Oarchitectureisthemostfundamental.
Inordertomeet the I/Orequirement andchallenge,alongwithourresearchtask ofa high performance parallel computing system, this paper is presenting our theoreticalstudy of I/O architectures, from which make it possible the high performance andscalability in terms of I/O architecture level. Meantime, I/O implementationmechanisms is focused on this paper, including technologies such as I/O-includedmemory consistency model and its implementation, intelligent I/O control, hybridstorage and transactional storage management, so as to promote I/O performance andavailability.
Themainworkandinnovativepointsofthispaperareasfollows.
1. I/Orestrictedparallelspeedup model
Current parallel I/O performance analysis lacks scientific theoretical models tosupporttheI/Oarchitecturedesign.ThepaperstudiestheimpactofI/Oworkloadonthescalability of parallel computing systems and proposes the I/O restricted parallelspeedupmodel.Basedonthismodel,whichcanbeusedtoguideI/Oarchitecturedesign,a scalable parallel I/O architecture for HPC is presented. Moreover, the paper analyzesseveral strategies for improving the system scalability, which serve as the basis forfurtherstudy.
2. I/O-includedgeneralmemoryconsistencymodelandimplementingtechnology
As for the consistency problem of shared memory systems with global DMAoperations,thepaperdefinestheconceptof I/O-includedgeneral program.Basedontheconcept, the paper studies the general memory consistency model, builds the generalsequence consistency model, general release consistency model and general scopeconsistency model. Using general scope memory consistency model, the paper designsand implements the CC-NUMA Cache Coherence protocol with global DMA and theglobal shared parallel I/O architecture at the hardware level. The experiment resultsshow that the I/O bandwidth and scalability of the system perform fairly well. Theactual parallel I/O bandwidth reaches 20.2 GB/s, and scales well with the number ofsystemprocesses.
3. IntelligentI/Oschedulealgorithmbasedonreinforcementlearning
To improve the I/O service efficiency of RAID and optimize the I/O performanceof parallel applications, the paper presents an intelligent I/O schedule algorithm,RL-scheduler, in RAID controllers based on reinforcement learning. RL-schedulerutilizes Q-learning strategy to implement a self-control and self-optimizing scheduler.The algorithm leverages the scheduling equity, disk seeking time and the I/O accessefficiency of MPI applications. Furthermore, the proposed interleaving organization ofmultiple Q-tables improves the efficiency of the Q-table updating. The experimentresults show that, on a large-scale parallel system with multiple parallel applications,RL-scheduler shortens the average I/O waiting time of parallel applicationsconsiderably. Thus increases the practical I/O bandwidth, and improves the systems’scalability.
4. Hybrid storagemanagement algorithmtosupport transactionsemantics
To address the requirement and challenge posed by HPC, the paper combines theidea of transactional storage management and hybrid storage acceleration, andintroduces an electro-magnetic hybrid storage management algorithm to supporttransaction semantics. A token-based protocol is designed to cope with the conflictsbetween I/O transactions and an adaptive logical partition algorithm is proposed tomanage Solid State Disk (SSD) storage. Simulation results show that theelectro-magnetic hybrid storage system can deal with transactions with varied accesspattern elegantly and effectively improve SSD hit rate, hide the overhead of versionmanagement and conflict detection. Both the I/O performance and availability aresignificantlyimproved.
引文
[1] TOP500Supercomputers.http://www.top500.org/,2009.
[2] Los Alamos National Laboratory. High-Performance Computing: RoadRunner.http://www.lanl.gov/roadrunner/,2009
[3] CarnegieMellonUniversity.PetascaleDataStorage.http://www.pdl.cmu.edu/PDSI/,2009.
[4] Santa Cruz University of California. Storage Systems Research Center.http://www.ssrc.ucsc.edu/proj/ceph.html,2009.
[5] Karsten Schwan. Petascale I/O for High End Computing. HEC FSIO Research andDevelopmentConference/HECURAFSIOPIMeeting'07.VA.Stafford,USA,2007,1-5.
[6] Andy Hospodor, Ethan L. Miller. Interconnection architectures for petabyte-scalehigh-performance storage systems. In Proceedings of the 21st IEEE / 12th NASA GoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,273-281.
[7] Sage A. Weil, Kristal T. Pollack, Scott A. Brandt, et al. Dynamic metadata management forpetabyte-scale file systems. In Proceedings of the 2004 ACM/IEEE Conference onSupercomputing(SC'04).Pittsburgh,PA,2004,1-7.
[8] Andrew Leung, Minglong Shao, Timothy Bisson, et al. High-Performance Metadata Indexingand Search in Petascale Data Storage Systems. Proceedings of the SciDAC 2008Conference.2008,1-12.
[9] SwapnilV.Patil,GarthA.Gibson,SamLang,etal.GIGA+:ScalableDirectoriesforSharedFile Systems. Proceedings of the 2nd international Petascale Data Storage Workshop(PDSW'07)heldinconjunctionwith Supercomputing'07.Reno,NV,2007,1-4.
[10] Oak Ridge National Laboratory/Lustre Center. Peta-Scale I/O with the Lustre File System.http://www.sun.com/offers/details/Peta-Scale_wp.html,2009.
[11] Andrew Leung, Minglong Shao, Timothy Bisson, et al. Spyglass: Fast, Scalable MetadataSearch for Large-Scale Storage Systems. Proceedings of the 7th USENIX Conference onFileandStorageTechnologies(FAST'09).2009,1-6.
[12] P.Caulk.TheDesignofaPetabyteArchiveandDistributionSystemfortheNASAECSProject. in Proceedings of the 4th NASA Goddard Conference on Mass Storage SystemsandTechnologies.CollegePark,MD,1995,1-3.
[13] J. Behnke, T. H. Watts. EOSDIS Petabyte Archives: Tenth Anniversary. in Proceedings ofthe 22nd IEEE/13th NASA Goddard Conference on Mass Storage Systems andTechnologies.Monterey,CA,2005,1-4.
[14] StandardLinearAcceleratorCenter.http://www.slac.stanford.edu/spires/experiments/online_exp.shtml,2009.
[15] CERN.http://press.web.cern.ch/press/PressReleases/Releases2008/PR17.08E.html,2008.
[16] A.T.Wong,L.Oliker,W.T.C.Kramer,etal.ESP:ASystemUtilizationBenchmark.inProceedingsofSupercomputing2000.Dallas,TX,2000,1-5.
[17] DOE National Nuclear Security Administration and the DOD National Security Agency.StatementofWork:SGSFileSystem.2001,1-4.
[18] E. L. Miller, R. H. Katz. Input/Output Behavior of Supercomputing Applications. inProceedingsofSupercomputing1991.Albuquerque,NM,1991,1-3.
[19] ASCIPurpleRFP.http://www.llnl.gov/asci/platforms/purple/rfp.
[20] ScalableIO:CurrentProjectsPDSISciDAC.http://www.cs.sandia.gov/Scalable_IO/.
[21] Schaller,RobertR.Moore'slaw:past,presentandfuture.IEEESpectrum,1997,14(6):52-59.
[22] Milo Polte, Jiri Simsa, Garth Gibson. Comparing Performance of Solid State Devices andMechanical Disks. Proceedings of the 3rd Petascale Data Storage Workshop held inconjunctionwithSupercomputing'08.Austin,TX. 2008,1-6.
[23] George VandenBerghe. I/O Trends at NOAA/NCEP. Thirteenth Workshop on Use of HighPerformanceComputinginMeteorology.UnitedKingdom,2008,1-10.
[24] http://en.wikipedia.org/wiki/IBM_Roadrunner,2009
[25] ErezZadok.FileSystemTracing,Replaying,Profiling,andAnalysisonHECSystems.HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '08. VA.Stafford,USA,2008,1-9.
[26] FengWang,QinXin,BoHong,etal. Filesystemworkloadanalysisforlarge scalescientificcomputing applications. In Proceedings of the 21st IEEE / 12th NASA GoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,139-152.
[27] T.M.Madhyastha, DanielA.Reed.Learningto classifyparallelinput/outputaccesspatterns.inIEEETransactionsonParallelandDistributedSystems, 2002,13(8):802-813.
[28] R. Ross, D. Nurmi, A. Cheng, et al. A case study in application I/O on Linux clusters. InProceedingsofSuperComputing(SC2001).Denver,CO,USA,2001,1-17.
[29] Roth. P. C. Characterizing the I/O behavior of scientific applications on the Cray XT. InProceedings of the 2nd international Workshop on Petascale Data Storage: Held inConjunctionwithSupercomputing'07.PDSW'07.NewYork,NY,2007,50-55.
[30] http://www.highproductivity.org/,2008
[31] Garth Gibson, Bianca Schroeder, Joan Digney, et al. Failure Tolerance in PetascaleComputers[J].VolumeonSoftwareEnablingTechnologiesforPetascaleScience.2007,1-6.
[32] BiancaSchroeder,GarthA.Gibson.UnderstandingDiskFailureRates:WhatdoesanMTTFof 1,000,000 hours mean to you? Proceedings of the 5th USENIX Conference on File andStorageTechnologies(FAST'07)BestPaperAward.SanJose,CA,2007,1-9.
[33] BiancaSchroeder,Garth A.Gibson.Understanding FailuresinPetascaleComputers. JournalofPhysics:ConferenceSeries,78.2007,1-6.
[34] QiongLi,EnqiangZhou,ZhongLiu,YufengGuo,XuejunYang. High-PerformanceStoragefor DSM System Using Global Cache, the 10th IEEE International Conference on HighPerformanceComputing andCommunications,2008,971-975.
[35] DrorG. Feitelson, Peter F. Corbett, SandraJohnson Baylor, etal. ParallelI/O Subsystems inMassivelyParallelSupercomputers.IEEEParallel&distributedtechnology,1995,1-6.
[36] R. Jain, J. Werth, J. C. Browne. Input/output in parallel and distributed computersystemsKluwerAcademicPublishers,1996.
[37] JohnM.May.ParallelI/OforHighPerformanceComputingAcademicpress,2001.
[38] M. R. Fahey, J. Larkin. Guidelines for Efficient Parallel I/O on the Cray XT3/XT4.Proceedingsofthe49thCrayUserGroup(CUG07).2007,1-5.
[39] Weikuan Yu, Vetter, J. S, et al. Performance characterization and optimization of parallelI/O on the CrayXT. IEEE International Symposiumon Parallel and Distributed Processing,IPDPS2008. 2008,1-11.
[40] Rob Ross(ANL), Jose Moreira(IBM), Kim Cupps(LLNL), et al. Parallel I/O on the IBMBlueGene/LSystem.2007,1-5.
[41] Weikuan Yu, Sarp Oral, Jeffrey Vetter, et al. Efficiency Evaluation of Cray XT Parallel IOStack.2007,1-6.
[42] S.R. Alam, R. F. Barrett, M. R. Fahey, et al. Cray XT4: An Early Evaluation for PetaScaleScientific Simulation. in ACM/IEEE conference on High Performance Networking andComputing(SC07).Reno,NV,2007,1-9.
[43] M. Fahey, J. Larkin, J. Adams. I/O Performance on a Massively Parallel Cray XT3/XT4. in22nd IEEE International Parallel & Distributed Processing Symposium (IPDPS'08). Miami,FL,2008,1-8.
[44]庞征斌,李琼,李永进,张峻,徐炜遐,CC-NUMA系统分布共享I/O的数据一致性维护,计算机研究与发展,2007,44(suppl.):226-232.
[45] Shane Canon, H. Sarp Oral. A Center-Wide File System using Lustre. Oak Ridge NationalLaboratory,2006,1-4.
[46] James Laudon, Daniel Lenoski. The SGI Origin: a ccNUMA highly scalable server.Proceedings of the 24th Annual International Symposium on Computer Architecture(ISCA-97).NewYork,NY,USA,1997,241-251.
[47] SGI-Products:ServersandSupercomputers:SGIAltixFamily.http://www.sgi.com/products/servers/altix/,2008
[48] Windsor W. Hsu, Alan Jay Smith. The performance impact of I/O optimizations and diskimprovements.IBMJournalofresearchanddevelopment,2004,48(2):1-7.
[49] XiaoningDing;SongJiang;FengChen. A Buffer Cache Management Scheme ExploitingBothTemporalandSpatialLocalities[J].ACM TransactionsonStorage.2007,1-20.
[50] DavidLeong.CollaborativeObjectCachingforHeterogeneousOSDClusters.Proceedingsofthe2007ComputerScienceandITEducationConference.2007,426-436.
[51] M. Korupolu, M. Dahlin. Coordinated placement and replacement for large-scale distributedcaches.IEEETransactionsonKnowledgeandDataEngineering,2002,1317-1329.
[52] Wei-keng Liao, K. Coloma, A. Choudhary, et al. Cooperative Client-Side File Caching forMPI Applications[J]. International Journal of High Performance Computing Applications2007.2007,144-154.
[53] SongJiang, KeiDavis, Xiaodong Zhang.Coordinated MultilevelBufferCache Managementwith Consistent Access Locality Quantification. IEEE transactions on computers,2007,56(1):1-6.
[54] He Xubin, Ben Ou Li. A unified multiple-level cache for high performance storage systems.InternationalJournalofHighPerformanceComputingandNetworking2007. 2007,1-4.
[55] Guang R. Gao Brian Lucas. A High Throughput Massive I/O Storage Hierarchy forPETA-scale High-end Architectures. HEC FSIO Research and DevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-9.
[56] Jiang Song, Zhang Xiaodong. Coordinating Existing Prefetching Strategies for Multi-levelStorage Systems. HEC FSIO Research and Development Conference/HECURA FSIO PIMeeting'08.VA.Stafford,USA,2008,1-8.
[57] Z. Chen, Y. Zhang, H. Zhou, Y., Scott, et al. Empirical evaluation of multi-level buffercache collaboration for storage systems. ACM SIGMETRICS: Performance EvaluationReview,2005,33(1):1-7.
[58] Florin Isaila, Guido Malpohl. Integrating Collective I/O and Cooperative Caching into the"Clusterfile"ParallelFileSystem.ICS’04.Malo,France,2004,1-4.
[59] Yong Chen. Hiding I/O Latency with Pre-execution Prefetching for Parallel Applications.SC2008.Austin,Texas,2008,1-2.
[60] Surendra Byna. Parallel I/O Prefetching Using MPI File Caching and I/O Signatures.SC2008.Austin,Texas,2008,1-4.
[61] C. K. Yang, T. Mitra, T. Chiueh. A Decoupled Architecture for Application-Specific FilePrefetching.inFreenixTrackofUSENIX2002AnnualConference.2002,1-4.
[62] S. Jiang, X. Zhang. STEP: Sequentiality and Thrashing Detection Based Prefetching toImprove Performance of Networked Storage Servers. in 27th International Conference onDistributedComputingSystems(ICDCS'07).2007,1-4.
[63] YongChen, Surendra Byna, Xian-He Sun. DataAccessHistoryCache andAssociated DataPrefetchingMechanisms.SC07.Reno,Nevada,USA,2007,1-5.
[64] Xiaosong Ma.Coordinating ExistingPrefetching StrategiesforMulti-levelStorage Systems.HEC FSIO Research and Development Conference/HECURA FSIO PI Meeting '08. VA.Stafford,USA,2008,1-6.
[65] XiaosongMa.Application-adaptiveI/OStackforData-intensiveScientificComputing.HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '07. VA.Stafford,USA,2007,1-4.
[66] AveryChing Wei-keng Liao. Using MPI File Caching to Improve Parallel Write PerformanceforLarge-ScaleScientificApplications.SC07.Reno,Nevada,USA,2007,1-6.
[67] Wei-keng Liao. Scalable I/O Middleware and File System Optimizations forHigh-Performance Computing. HEC FSIO Research and DevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-7.
[68] Mahmut Kandemir, Alok Choudhary. Compiler-Directed I/O Optimization. Proceedings oftheInternationalParallelandDistributedProcessingSymposium(IPDPS.02).2002,1-10.
[69] A. Acharya, M. Uysal, J. Saltz. Active disks: Programming model, algorithms andevaluation.InProc.oftheACMASPLOSConference.1998,81-91.
[70] K. Keeton, D. A. Patterson, J. M. Hellerstein. A case for intelligent disks (IDISKs). InSIGMODRecord,Vol.24,No.7,1998,42-52.
[71] E. Riedel, G. Gibson, C. Faloutsos . Active storage for large-scale data mining andmultimedia.InProc.ofthe24thInt.Conf.VeryLargeDataBases(VLDB).1998,62-73.
[72] P. M. Kogge, J. B. Brockman. Processing in memory: Chips to petaflops. In Workshop onMixingLogicandDRAM:ChipsthatComputeandRemember(atISCA'97).1997,1-6.
[73] Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau, Lakshmi N. Bairavasundaram.Semantically-smart disk systems: Past, present, and future. ACM SIGMETRICSPerformance Evaluation Review on Design, implementation, and performance of storagesystems, 2006,33(4):29-35.
[74] Ma Xiaonan, Reddy A. L. N. MVSS.An active storage architecture. IEEE Trans on ParallelandDistributedSystems,2003, 14(10):993-1005.
[75] G. A. Gibson, D. F. Nagle, K. Amiri, et al.File server scaling with network-attached securedisks. In Proc. Of the 1997 ACM SIGMETRICS International Conference on MeasurementandModelingofComputerSystems.1997,272-284.
[76] M. Mesnier, G. Ganger, E. Riedel. Object-based storage. IEEE Communications Magazine,2005,.41(8):84-90.
[77] Azagury A, Dreizin v, Factor M. Towards an object store. In: Proc of the 20th IEEE/11thNASAGoddardConferenceonMassStorageSystemsandTechnologies(MSST2003).2003,165-176.
[78] Sivathanu M,Arpaci-Dusseau A.C,Arpaci-Dusseau R.H.Evolving RPC foractive storage.ACM SIGPLAN Notices SESSION on Communication abstractions and optimizations,2002, 37(10):264-276.
[79] H. Lim, V. Kapoor, C. Wighe, et al. Active disk file system: A distributed, scalable filesystem. In Proc. of the 18th IEEE Symposium on Mass Storage Systems and Technologies.SanDiego,2001,101-115.
[80] Joel C. Wu, Scott A. Brandt. The design and implementation of AQuA: an adaptive qualityof service aware object-based storage device. In Proceedings of the 23rd IEEE 14th NASAGoddard Conference on Mass Storage Systems and Technologies. College Park, MD, 2006,209-218.
[81] The advantages of object-based storage -secure, scalable, dynamic storage devices.http://www.Seagate.com,2008
[82] D. Nagle, D. Serenyi, A. Matthews. The panasas activescale storage cluster - Deliveringscalable high bandwidth storage. Proceedings of the 2004 AGM/IEEE Conference onSupercomputing(SC2004).2004,-10.
[83] O. Rodeh, A. Teperman. zFS - A Scalable distributed File System using Object Disks. InProceedings of the IEEE Mass Storage Systems and Technologies Conference. San Diego,CA,USA,2003,207-218.
[84] Feng Wang, Scott A. Brandt, Ethan L. Miller, et al. OBFS: A file system for object-basedstorage devices. In Proceedings of the 21st IEEE / 12th NASA Goddard Conference onMassStorageSystemsandTechnologies.CollegePark,MD,2004,283-300.
[85] Scott A. Brandt Sage A.Weil, Ethan L. Miller, Darrell D. E. Long, et al. Ceph: A scalable,high-performance distributed file system. In Proceedings of the 7th Symposium onOperatingSystemsDesignandImplementation(OSDI).Seattle,WA,2006,1-6.
[86] Sage A. Weil, Scott A. Brandt, Ethan L. Miller, et al. CRUSH: Controlled, scalable,decentralized placement of replicated data. In Proceedings of the 2006 ACM/IEEEConferenceonSupercomputing(SC'06).Tampa,FL,2006,1-7.
[87] PeteWyckoff. ApplicabilityofObject-Based Storage DevicesinParallelFile Systems. HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '07. VA.Stafford,USA,2007,1-9.
[88] Brent Welch, Garth Gibson. Managing scalability in object storage systems for HPC Linuxclusters. In Proceedings of the 21st IEEE / 12th NASA Goddard Conference on MassStorageSystemsandTechnologies.2004,433-445.
[89] P.Braam.Thelustrestoragearchitecture.ClusterFileSystem,Inc,http://www.csis.hku.hk/cluster2003/presentation/vt/lustre.pdf,2007.
[90] E. J. Felix, K. Fox, K. Regimbal, etal.Active storage processing in a parallelfile system. InProc. of the 6th LCI International Conference on Linux Clusters: The HPC Revolution.2006,1-7.
[91] Juan Piernas. Evaluation of Active Storage Strategies for the LustreParallel File System. SC'07.Reno,Nevada,USA,2007,1-10.
[92] MitchellT.M.MachineLearning.McGraw-HillScience/Engineering/Math,1997,1-7.
[93] Sage Weil, Andrew Leung, Scott A. Brandt, et al. RADOS: A Fast, Scalable, and ReliableStorage Service for Petabyte-scale Storage Clusters. Proceedings of the ACM PetascaleDataStorageWorkshop2007(PDSW07).2007,1-5.
[94] Milo Polte, Jiri Simsa, Wittawat Tantisiriroj, et al. Fast Log-based Concurrent Writing ofCheckpoints. Proceedings of the 3rd Petascale Data Storage Workshop held in conjunctionwithSupercomputing'08.Austin,TX,2008,1-13.
[95] Eduardo Pinheiro, Wolf-Dietrich Weber, Luiz Andre Barros.Failure Trends in a Large DiskDrivePopulation.InFAST'07.SanJose,CA,2007,1-9.
[96] JonG.Elerath, MichaelPecht. Enhanced ReliabilityModeling ofRAID Storage Systems. InDSN'07.Edinburgh,UK,2007,1-7.
[97] Remzi Arpaci-Dusseau. Formal Failure Analysis for Storage Systems. HEC FSIO Researchand Development Conference/HECURA FSIO PI Meeting '07. VA. Stafford,USA, 2007,1-5.
[98] Qin Xin, Ethan L. Miller, Thomas J. E. Schwarz. Evaluation of distributed recovery inlarge-scale storage systems. In Proceedings of the 13th IEEE International Symposium onHighPerformanceDistributedComputing(HPDC).Honolulu,HI,2004,172-181.
[99] Soila Pertet, Rajeev Gandhi, Priya Narasimhan. Fingerpointing Correlated Failures inReplicated Systems. USENIX Workshop on Tackling Computer Systems Problems withMachineLearningTechniques(SysML).Cambridge,MA,2007,1-8.
[100] Muthian Sivathanu, Vijayan Prabhakaran, Andrea C. Arpaci-Dusseau, et al. ImprovingStorageSystemAvailabilitywithD-GRAID.InFAST'04.SanFrancisco,CA,2004,1-6.
[101] Lei Tian, Dan Feng, Hong Jiang, et al. PRO: A Popularity-based Multi-threadedReconstruction Optimization for RAID-Structured Storage Systems. In FAST '07. San Jose,CA,2007,1-9.
[102] Kiron Vijayasankar, Gopalan Sivathanu, Swaminathan Sundararaman. Exploiting Type-AwarenessinaSelf-RecoveringDisk.InStorageSS'07.Alexandria,Virginia,2007,1-8.
[103] Kevin Greenan, Ethan Miller, Thomas Schwarz, et al. Disaster Recovery Codes: IncreasingReliability with Large-Stripe Error Correction Codes. In StorageSS '07. Alexandria,Vir-ginia,2007,1-11.
[104] Haryadi S. Gunawi, Vijayan Prabhakaran, Swetha Krishnan, et al. Improving File SystemReliabilitywithI/OShepherding.InSOSP'07.Stevenson,WA,2007,1-7.
[105] http://msdn.microsoft.com/en-us/magazine/cc163388.aspx,2009
[106] http://tservice.net.ru/~s0mbre/old/?section=projects&item=pohmelfs, 2009
[107] http://commons.apache.org/transaction/file/index.html,2009
[108] B. Hong F. Wang, Q. Xin, S. A. Brandt, E. L. Miller, et al. File System Workload AnalysisFor Large Scale Scientific Computing Applications. in Proceedings of the 21st IEEE/12thNASAGoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,1-3.
[109]卢凯.面向并行科学计算应用的并行文件系统技术研究,博士论文,国防科学技术大学,1999.
[110] DanielA.Reed. Workshop on the Road Map forthe Revitalization ofHigh End Computing.http://www.cra.org/reports/supercomputing.pdf,2003.
[111] Nitin Agrawal, Vijayan Prabhakaran, Ted Wobber, et al. Design tradeoffs for SSDperformance.InUSENIXAnnualTechnicalConference.Boston,MA,2008,1-4.
[112] Dushyanth Narayanan, Eno Thereska, Austin Donnelly. Migrating enterprise storage toSSDs:analysisoftradeoffs. TechnicalMicrosoftResearchCambridge,UK,2008.
[113] DataStorageTechnologies-Transitiontosolidstate.http://www.ru.is/kennarar/andri/nyti/papers2007/DST.pdf,2007.
[114] Kang J, Jo H, Kim J. A superblock-based flash translation for NAND flash memory. NewYork,NY,USA,2006,1-5.
[115] B. Banister, J. Bowen. High Performance QFS with Solid State Disk Metadata Storage.TechnicalTexasMemorySystems,Inc,2005.
[116] DavidDriggers.TheTop5TrendsinHighPerformanceComputingfor2009.http://www.wwpi.com/systems/systems/6663-the-top-5-trends-in-high-performance-computing-for-2009,2009.
[117] Jing Li, Patrick Ndai, Ashish Goel, et al. An Alternate Design Paradigm for RobustSpin-Torque Transfer Magnetic RAM (STT MRAM) from Circuit/Architecture Perspective.14thAsiaandSouthPacificDesignAutomationConference.2009,1-6.
[118] Kato, Yoshihisa Tanaka, Hiroyuki Isogai, et al. Embedded FeRAMChallengesin the 65-nmTechnology Node and Beyond. Applications of ferroelectrics, 2006. isaf '06. 15th ieeeinternationalsymposium.2006,1-4.
[119] Jin Hyuk Yoon, Eyec Hyun Nam, Yoon Jae Scong, et al. Chameleon: A High PerformanceFlash/FRAM Hybrid Solid State Disk Architecture. Computer Architecture Letters,20071-7.
[120] T. Rueckes, K. Kim, E. Joselevich. Carbon Nanotube-based Nonvolatile Random AccessMemoryForMolecularComputing.Science,2000,298:94-97.
[121] Mustafa Uysal, Arif Merchant, Guillermo A. Alvarez. Using MEMS-based storage in diskarrays. Proceedings of the 2nd USENIX Conference on File and Storage Technologies. SanFrancisco,CA,2003,1-9.
[122] R. F. FREITAS, W. W. WILCKE. Storage-class memory. The next storage systemtechnology.IBMjournalofresearchanddevelopment,2008, 52:1-5.
[123] G. W. Burr, B. N. Kurdi, J. C. Scott, et al. Overview of candidate device technologies forstorage-class memory. (IBM Journal of Research and development) Storage TechnologiesandSystems,2008, 52:4-5.
[124] Winfried Wilcke. Flash and Storage Class Memories: Technology Overview & SystemsImpact.HECFSIOResearchandDevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-8.
[125] An-I A. Wang, Peter Reiher, Gerald J. Popek, et al. Conquest: better performance through adisk/persistent-RAM hybrid file system. In Proceedings of the 2002 USENIX AnnualTechnicalConference.Monterey,2002,15-28.
[126] Ethan L. Miller. HeRMES: High-Performance Reliable MRAM-Enabled Storage. 8th IEEEWorkshop on Hot Topics in Operating Systems (HOTOS-VIII). Schloss Elmau, Germany,2001,83-87.
[127] I. Ari, M. Gottwals, D. Henze. SANBoost: Automated SAN-Level Caching in Storage AreaNetworks. In Proceedings of the 13th IEEE International Conference on AutonomicComputing(ICAC).2004,164-171.
[128] M.Casey.SANCache:SSDintheSAN.TechnicalSoliddata,Inc,2000.
[129] Gurumurthi S, Sivasubramaniam A, Kandemir M. DRPM: Dynamic Speed Control forPower Management in Server Class Disks. In Proc. of the 30th annual internationalsymposiumonComputerarchitecture.SanDiego,California,2003,169-181.
[130] Gurumurthi S, Zhang J, Sivasubramaniam A. Interplay of Energy and Performance for DiskArrays Running Transaction Processing Workloads. In Proc. of 2003 IEEE InternationalSymposium on Performance Analysis of Systems and Software. Austin, Texas; USA,2003,123-132.
[131] Adam Roth. Energy Conservation for Real-Time Disk Systems with I/O Burstiness. IEEEInt'lWorkshopNextGenerationAutonomousStorageandHighPerformanceComputing.St.Thomas,VirginIslands,2008,1-4.
[132] Mais Nijim. An Adaptive Energy-Conserving Strategy for Parallel Disk Systems. in theProceedings of the 12th IEEE International Symposium on Distributed Simulation and RealTimeApplications(DS-RT'08).Vancouver,BritishColumbia,Canada,2008,1-5.
[133] CharlesWeddle.PARAID:AGear-ShiftingPower-AwareRAID,1-7.
[134] Eduardo Pinheiro. Energy Conservation Techniques for Disk Array-Based Servers. ICS'04.Malo,France,2004,1-8.
[135] Seung Woo Son. EnergyAware Data Prefetching for MultiSpeed Disks. CF'06. Ischia, Italy,2006,1-3.
[136] LuisUseche.EXCES:EXternalCachinginEnergySavingStorageSystems,1-6.
[137] JianhuiYue.AnEnergy-Oriented Evaluation ofBufferCacheAlgorithmsUsingParallelI/OWorkloads. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2008, 19(11):1-9.
[138] DavidE.Culler,JaswinderPalSingh,AnoopGupta.并行计算机体系结构:硬件与软件结合的设计与分析,机械工业出版社,2002.
[139] Mark D. Hill, Anne E. Condon, Manoj Plakal, et al. A System-Level SpecificationFramework for I/O Architectures. the 11th Annual Symposium on Parallel Algorithms andArchitectures(SPAA).1999,1-6.
[140] John Wilkes, Richard Golding, Carl Staelin, et al. The HP AutoRAID heirarchical storagesystem.IEEETransactiononcomputersystems, 1996, 14(1):108-136.
[141] E. Anderson, M. Kallahalla, S. Spence, et al. Quickly finding near-optimal storage systemdesigns.http://hpl.hp.com/research/ssp/papers/ergastulum-paper.pdf,2005.
[142] Mohamed N. Bennani, Daniel A. Menasce. Resource allocation for autonomic data centersusing analytic performance models. In The 2nd IEEE International Conference onAutonomicComputing(ICAC-05).Seattle,WA,2005,229-240.
[143] Hewlett-PackardCompany.Adaptiveenterprise.http://www.hp.com/products1/promos/adaptive/enterprise/us/adaptive/enterprise.html,2005.
[144] M. Wang. Black-Box Storage Device Modeling with Learning:[PhD dissertation]. CarnegieMellonUniversity,2006.
[145] Engin Ipek, Onur Mutlu, Jose F. Martnez, et al. Self-Optimizing Memory Controllers. AReinforcement Learning Approach, Intl. Symp. on Computer Architecture (ISCA). Beijing,China,2008,1-8.
[146] Yu Zhang, Bharat Bhargava. Self-Learning Disk Scheduling. IEEE Transactions onknowledgeanddataengineering,2009,21(1):1-9.
[147] D. L. Martens, M. J. Katchabaw. Optimizing System Performance through Dynamic DiskScheduling Algorithm Selection,WSEAS Trans. Information Science and Applications,2006,1-7.
[148] M. Stillger, G. Lohman, V. Markl, et al. LEO-DB2'S Learning Optimizer. Proc. 27thinternationalconferenceof VeryLargeDataBases(VLDB).2001,1-5.
[149] K. Shen, M. Zhong, C. Li. I/O System Performance Debugging Using Model-DrivenAnomaly Characterization. Proc. Fourth Usenix Conf. File and Storage Technologies(FAST'05).2005,1-7.
[150] J. Wildstrom, P. Stone, E. Witchel, et al. Machine Learning for On-Line HardwareReconfiguration.Proc.20thInt'lJointConf.ArtificialIntelligence(IJCAI'07).2007,1-9.
[151] http://www.cs.waikato.ac.nz/ml/weka/,2008.
[152] William W. Cohen. Fast effective rule induction. In Proceedings of the 12th InternationalConferenceonMachineLearning(ICML-95).1995,15-123.
[153]舒继武,薛巍,付长冬.网络存储系统与技术的现状与发展趋势.中国计算机学会通信,2004,1-5.
[154] Maurice Herlihy, J. EliotMoss. TransactionalMemory:ArchitecturalSupportfor Lock-FreeData Structures. Proceedings of the 20th Annual International Symposium on ComputerArchitecture.1993,289--300.
[155] C. Scott Ananian, Krste Asanovic, Bradley C. Kuszmaul, et al. Unbounded TransactionalMemory. Proceedings of the Eleventh International Symposium on High-PerformanceComputerArchitecture.2005,316--327.
[156] Lance Hammond, Vicky Wong, Mike Chen, et al. Transactional Memory Coherence andConsistency. Proceedings of the 31st Annual International Symposium on ComputerArchitecture.2004,102.
[157] Kevin E. Moore, Jayaram Bobba, Michelle J. Moravan, et al. LogTM: Log-basedTransactional Memory. Proceedings of the 12th International Symposium onHigh-PerformanceComputerArchitecture.2006,254--265.
[158] Luke Yen, Jayaram Bobba, Michael M. Marty, et al. LogTM-SE: Decoupling HardwareTransactional Memory from Caches. Proceedings of the 13th International Symposium onHigh-PerformanceComputerArchitecture(HPCA).2007,1-7.
[159] JayaramBobba,NeelamGoyal, Mark D. Hill, etal. TokenTM: EfficientExecution of LargeTransactions with Hardware Transactional Memory. Proceedings of the 35th InternationalSymposiumonComputerArchitecture.IEEEComputerSociety,2008.2008,127-138.
[160] Maurice Herlihy, Victor Luchangco, Mark Moir, et al. Software Transactional Memory forDynamic-SizedDataStructures. 2003,92--101.
[161] DaveDice,OriShalev,NirShavit.TransactionalLockingII.2006,1-5.
[162] Bratin Saha, Ali- Reza Adl-Tabatabai, Richard L. Hudson, et al. McRT-STM: a HighPerformance Software Transactional Memory System for a Multi-core Runtime. Proc. 11thACM SIGPLAN Symp. on Principles and Practice of Parallel Programming (PPoPP '06).2006,187-197.
[163] Arrvindh Shriraman, Sandhya Dwarkadas, Michael L. Scott. Flexible DecoupledTransactional Memory Support. Proceedings of the 35th Annual International SymposiumonComputerArchitecture.2008,1-6.
[164] Cao Minh, Martin Trautmann, Jae Woong Chung, et al. An Effective Hybrid TransactionalMemory System with Strong Isolation Guarantees. Proceedings of the 34th AnnualInternationalSymposiumonComputerArchitecture.2007,1-9.
[165] Dice Yossi Dave, Dan Nussbaum. Early Experience with a Commercial HardwareTransactional Memory Implementation. Proceedings of the 14th International ConferenceonArchitecturalSupportforProgrammingLanguagesandOperatingSystems.2009,1-11.
[166] http://www.tevero.no/products/tffs/index.htm.
[167] http://opensolaris.org/os/community/zfs.
[168] Zhihui Zhang, Kanad Ghose. hfs: A hybrid file system prototype for improving small fileandmetadataperformance.InProceedingsofEuroSys2007.2007,1-4.
[169] DataDirectNetwor.DDNProducts:S2A9550.http://www.ddn.com,.
[170] EMCCLARiiONCX4.http://www.emc.com,.
[171] http://www.netapp.com.
[172] B. Banister, J. Bowen. High Performance QFS with Solid State Disk Metadata Storage.TechnicalTexasMemorySystems,Inc,2005.
[173] BoHong,ScottABrandt,DarrellELong,EthanL.Miller,etal.UsingMEMS-BasedStorageinComputerSystems-DeviceModelingandManagement.2006,1-7.
[174] FengWang,BoHong, Scott A. Brandt, Darrell D. E. Long. Using MEMS-based Storage toBoostDiskPerformance.2005,1-9.
[175] Steven W. Schlosser, Jiri Schindler, Anastassia Ailamaki, et al. Exposing and exploitinginternalparallelisminMEMS-basedstorage2003.
[176] PaulMassiglia.TheRAIDbook,6thEDITION.
[177] Yashushi Saito, Svend Frflund, Alistair Veitch. FAB: Building distributed enterprise diskarraysfromcommoditycomponents.InProceedingsofthe11thInternationalConferenceonArchitectural Support for Programming Languages and Operating Systems (ASPLOS).2004,48-58.
[178] John L. Gustafson. Reevaluating Amdahl's Law. Communications of the ACM,1988,31(5):532-533.
[179] A. Moga, A. Gefflaut, M. Dubois. Hardware Versus Software Implementation of COMA.Proceedingsofthe1997InternationalConferenceParallelProcessing.1997,248-255.
[180] P. Keleher, S. Dwarkadas, A. L. Cox. TreadMarks: Distributed Shared Memoryon StandardWorkstations and Operating Systems. In Proceedings of the Winter 1994 USENIXConference.1994,115-131.
[181] OpenMP:Simple,Portable,ScalableSMPProgramming.http://www.openmp.org/drupal/ .
[182] Rajesh A. Bordawekar. Quantitative Characterization and Analysis of the I/O Behavior of aCommercial Distributed-Shared-Memory Machine. IEEE Trans. Parallel Distrib. Syst.2000,11(5):509--526.
[183] L.Lamport.HowtoMakeaMultiprocessorComputerThatCorrectlyExecutesMultiprocessPrograms.IEEETransactionsonComputers,1979,28(9):690-691.
[184] M. Dubois, C. Scheurich, F. A. Briggs. Memory Access Buffering in Multiprocessors. Proc.13thInt'lSymp.onComputerArchitecture.1986,434-442.
[185] K. Gharachorloo, D. Lenoski. Memory Consistency and Event Ordering in ScalableShared-Memory Multiprocessors. In Proceedings of the 17th annual InternationalSymposiumonComputerArchitecture.Seattle,Washington,1990,15-26.
[186] J. P. Singh L.Iftode, K. Li Scope Consistency. A Bridge Between Release Consistency andEntry Consistency. In Proceedings of the 8th Annual ACM Symposium on ParallelAlgorithmsandArchitectures.1996,1-9.
[187] B. N. Bershad, M. J. Zekauskas, W. A. Sawdon. The Midway Distributed Shared MemorySystem. InProceedingsofthe38thIEEEInt'lComputerConf(COMPCONSpring'93).1993,528-537.
[188] Z. Huang, C. Sun, M. Purvis. View-based Consistency for Distributed Shared Memory. InProceedings of the 4th World Multiconference on Systemics, Cybernetics and Informatics(SCI'2000).2000,1-7.
[189] Z. Huang, C. Sun, S. Cranefield. View-based Consistency and Its Implementation. InProceedings of the 1st IEEE/ACM International Symposium on Cluster Computing and theGrid.Brisbane,2001,74-81.
[190] Z. Huang, C. Sun, M. Purvis. View-Based Consistency and False Sharing Effect inDistributedSharedMemory.InOperatingSystemsReview,SIGOPS, 2001, 35(2):51-60.
[191] Kourosh Gharachorloo. Memory Consistency Models for Shared-Memory Multiprocessors.PhD.Dissertation,1995,1-47.
[192] Yang Xuejun, Dai Huadong. Operating System-centric Memory consistency model--threadconsistency model. The Fourth International Workshop on Advanced Parallel ProcessingTechnologies.2001,1-8.
[193]戴华东.并行分布操作系统共享存储管理与优化技术研究,博士论文,国防科学技术大学,2002.
[194]胡伟武.共享存储系统结构,高等教育出版社,2001.
[195] Scheurich C, Dubois M. Correct Memory Operation of Cached-Based Multiprocessors.Proceeding of the 14th Anuual International Symposium on Computer Architecture. 1987,1-6.
[196] John L. Hennessy, David A. Patterson.计算机体系结构-量化研究方法机械工业出版社,2002.
[197]吴俊敏,杨超,陈国良.存储模型仿真器的设计与实现.计算机研究与发展,2004,42(3):394-403.
[198]庞征斌.基于SMP的CC-NUMA类大规模系统中Cache一致性协议研究与实现,博士论文,国防科学技术大学,2007.
[199] David E. Culler, Jaswinder Pal Singh, Anoop Gupta. Parallel Computer Architecture: AHardware/SoftwareApproach(SecondEdition)MorganKaufmannPress,1996.
[200] NAS Parallel Benchmarks I/O Version 3.2. NASA Advanced Supercomputing (NAS)Division,2005.
[201] P. Scheuermann, G. Weikum, P. Zabback. Data partitioning and load balancing in paralleldisksystems.VLDBJournal:VeryLarge DataBases,1998,7(1):48-66.
[202] IORBenchmark.http://www.llnl.gov/asci/purple/benchmarks/limited/ior,.
[203]莫则尧,许林宝,张宝琳.二维等离子体模拟粒子云网格方法的并行计算与性能分析.计算物理,1998,16(5):496-504.
[204] R. K. Sahoo H. Yu, C. Howson. High Performance File I/O for the Blue Gene/LSupercomputer,2007,1-6.
[205] Shenze Chen, John A. Stankovic, James F. Kurose, et al. Performance Evaluation of TwoNewDiskSchedulingAlgorithmsforReal-TimeSystem,1990.
[206] IozoneFilesystemBenchmark.http://www.iozone.org/src/current/iozone3_263.tar,2008.
[207] Elizabeth J. O'Neil;Patrick E. O'Neil, Gerhard Weikum. The lru-k page replacementalgorithm for database disk buffering. In Proceedings of the 1993 ACM SIGMODinternationalconferenceonManagementofdata.NewYork,NY,USA,1993,297-306.
[208] Donghee Lee, Jongmoo Choi, Jong-Hun Kim, et al. On the existence of a spectrum ofpolicies that subsumes the least recently used (LRU) and least frequently used (LFU)policies. In Proceedings of the 1999 ACM SIGMETRICS International conference onMeasurementandModelingofComputerSystems.1999,134-143.
[209] Y. Smaragdakis, S. Kaplan, P. Wilson. EELRU: Simple and Effective Adaptive PageReplacement.InProceedingsofthe1999ACMSIGMETRICSConferenceonMeasurementandModeling ofComputerSystems.1999,122-133.
[210] Yuanyuan Zhou, James Philbin, Kai Li. The multi-queue replacement algorithm for secondlevel buffer caches. In Proceedings of the General Track: 2002 USENIX Annual TechnicalConference.Berkeley,CA,USA,2001,91-104.
[211] Song Jiang, Xiaodong Zhang. LIRS: An efficient low inter-reference recency setreplacement policy to improve buffer cache performance. In Proceedings of the ACMSIGMETRICS Conference on Measurement and Modeling of Computer Systems. 2002,1-12.
[212] Nimrod Megiddo, Dharmendra S. Modha. ARC: A self-tuning, low overhead replacementcache. In Proceedings of the 2nd USENIX Conference on File and Storage Technologies(FAST).2003,115-130.
[213] Pei Cao, Edward W. Felten, Anna R. Karlin, et al. Implementation and performance ofintegrated application-controlled file caching, prefetching, and disk scheduling. ACMTransactionsonComputerSystems, 1996,14(4):311-343.
[214] R. Hugo Patterson, Garth A. Gibson, Eka Ginting, et al. Informed prefetching and caching.In Proceedings of the fifteenth ACM symposium on Operating systems principles (SOSP).NewYork,NY,USA,1995,79-95.
[215]TheodoreJohnson,DennisShasha.2Q:A lowoverhead highperformance buffermanagementreplacement algorithm. In Proceedings of the 20th International Conference on Very LargeDataBases(VLDB).SanFrancisco,CA,USA,1994,439-450.
[216] Gideon Glass, PeiCao. Adaptive page replacement based on memoryreference behavior. InProceedings of the ACM SIGMETRICS international conference on Measurement andModelingofComputerSystems.1997,115-126.
[217] Chris Gniady, Ali Raza Butt, Y. Charlie Hu. Program-counter-based pattern classification inbuffer caching. In Proceedings of 6th Symposium on Operating System Design andImplementation(OSDI).2004,395-408.
[218] Jong Min Kim, Jongmoo Choi, Jesung Kim, et al. A low-overhead, high-performanceunified buffer management scheme that exploits sequential and looping references. In 4thSymposiumonOperatingSystemDesignandImplementation(OSDI).2000,1-6.
[219] Feng Zhou, Rob von Behren, Eric Brewer. AMP: Program context specific buffer caching.InProceedingsoftheUSENIXTechnicalConference.2005,1-8.
[220] Jongmoo Choi, Sam H. Noh, Sang Lyul Min, et al. An implementation study of adetection-based adaptive block replacement scheme. In Proceedings of the 1999 USENIXAnnualTechnicalConference.1999,239-252.
[221] Yifeng Zhu. Cluster-based storage systems with high scalability:[PhD Thesis]. UniversityofNebraska,2005.
[222] J. Choi, S. Cho, S. H. Noh, et al. Analytic Prediction of Buffer Hit Ratios. IEEElectronicsLetters,2000,36(1):10-11.
[223] R. H. Patterson, G. A. Gibson, E. Ginting, et al. Informed Prefetching and Caching. InProceedingsofthe15thSymposiumonOperatingSystemPrinciples.1995, 1-16.
[224] G. R. Ganger, B. L. Worthington, Y. N. Patt. The DiskSim simulation environment version2.0 reference manual. Technical Carnegie Mellon University / University of Michigan,1999.
[2] Los Alamos National Laboratory. High-Performance Computing: RoadRunner.http://www.lanl.gov/roadrunner/,2009
[3] CarnegieMellonUniversity.PetascaleDataStorage.http://www.pdl.cmu.edu/PDSI/,2009.
[4] Santa Cruz University of California. Storage Systems Research Center.http://www.ssrc.ucsc.edu/proj/ceph.html,2009.
[5] Karsten Schwan. Petascale I/O for High End Computing. HEC FSIO Research andDevelopmentConference/HECURAFSIOPIMeeting'07.VA.Stafford,USA,2007,1-5.
[6] Andy Hospodor, Ethan L. Miller. Interconnection architectures for petabyte-scalehigh-performance storage systems. In Proceedings of the 21st IEEE / 12th NASA GoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,273-281.
[7] Sage A. Weil, Kristal T. Pollack, Scott A. Brandt, et al. Dynamic metadata management forpetabyte-scale file systems. In Proceedings of the 2004 ACM/IEEE Conference onSupercomputing(SC'04).Pittsburgh,PA,2004,1-7.
[8] Andrew Leung, Minglong Shao, Timothy Bisson, et al. High-Performance Metadata Indexingand Search in Petascale Data Storage Systems. Proceedings of the SciDAC 2008Conference.2008,1-12.
[9] SwapnilV.Patil,GarthA.Gibson,SamLang,etal.GIGA+:ScalableDirectoriesforSharedFile Systems. Proceedings of the 2nd international Petascale Data Storage Workshop(PDSW'07)heldinconjunctionwith Supercomputing'07.Reno,NV,2007,1-4.
[10] Oak Ridge National Laboratory/Lustre Center. Peta-Scale I/O with the Lustre File System.http://www.sun.com/offers/details/Peta-Scale_wp.html,2009.
[11] Andrew Leung, Minglong Shao, Timothy Bisson, et al. Spyglass: Fast, Scalable MetadataSearch for Large-Scale Storage Systems. Proceedings of the 7th USENIX Conference onFileandStorageTechnologies(FAST'09).2009,1-6.
[12] P.Caulk.TheDesignofaPetabyteArchiveandDistributionSystemfortheNASAECSProject. in Proceedings of the 4th NASA Goddard Conference on Mass Storage SystemsandTechnologies.CollegePark,MD,1995,1-3.
[13] J. Behnke, T. H. Watts. EOSDIS Petabyte Archives: Tenth Anniversary. in Proceedings ofthe 22nd IEEE/13th NASA Goddard Conference on Mass Storage Systems andTechnologies.Monterey,CA,2005,1-4.
[14] StandardLinearAcceleratorCenter.http://www.slac.stanford.edu/spires/experiments/online_exp.shtml,2009.
[15] CERN.http://press.web.cern.ch/press/PressReleases/Releases2008/PR17.08E.html,2008.
[16] A.T.Wong,L.Oliker,W.T.C.Kramer,etal.ESP:ASystemUtilizationBenchmark.inProceedingsofSupercomputing2000.Dallas,TX,2000,1-5.
[17] DOE National Nuclear Security Administration and the DOD National Security Agency.StatementofWork:SGSFileSystem.2001,1-4.
[18] E. L. Miller, R. H. Katz. Input/Output Behavior of Supercomputing Applications. inProceedingsofSupercomputing1991.Albuquerque,NM,1991,1-3.
[19] ASCIPurpleRFP.http://www.llnl.gov/asci/platforms/purple/rfp.
[20] ScalableIO:CurrentProjectsPDSISciDAC.http://www.cs.sandia.gov/Scalable_IO/.
[21] Schaller,RobertR.Moore'slaw:past,presentandfuture.IEEESpectrum,1997,14(6):52-59.
[22] Milo Polte, Jiri Simsa, Garth Gibson. Comparing Performance of Solid State Devices andMechanical Disks. Proceedings of the 3rd Petascale Data Storage Workshop held inconjunctionwithSupercomputing'08.Austin,TX. 2008,1-6.
[23] George VandenBerghe. I/O Trends at NOAA/NCEP. Thirteenth Workshop on Use of HighPerformanceComputinginMeteorology.UnitedKingdom,2008,1-10.
[24] http://en.wikipedia.org/wiki/IBM_Roadrunner,2009
[25] ErezZadok.FileSystemTracing,Replaying,Profiling,andAnalysisonHECSystems.HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '08. VA.Stafford,USA,2008,1-9.
[26] FengWang,QinXin,BoHong,etal. Filesystemworkloadanalysisforlarge scalescientificcomputing applications. In Proceedings of the 21st IEEE / 12th NASA GoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,139-152.
[27] T.M.Madhyastha, DanielA.Reed.Learningto classifyparallelinput/outputaccesspatterns.inIEEETransactionsonParallelandDistributedSystems, 2002,13(8):802-813.
[28] R. Ross, D. Nurmi, A. Cheng, et al. A case study in application I/O on Linux clusters. InProceedingsofSuperComputing(SC2001).Denver,CO,USA,2001,1-17.
[29] Roth. P. C. Characterizing the I/O behavior of scientific applications on the Cray XT. InProceedings of the 2nd international Workshop on Petascale Data Storage: Held inConjunctionwithSupercomputing'07.PDSW'07.NewYork,NY,2007,50-55.
[30] http://www.highproductivity.org/,2008
[31] Garth Gibson, Bianca Schroeder, Joan Digney, et al. Failure Tolerance in PetascaleComputers[J].VolumeonSoftwareEnablingTechnologiesforPetascaleScience.2007,1-6.
[32] BiancaSchroeder,GarthA.Gibson.UnderstandingDiskFailureRates:WhatdoesanMTTFof 1,000,000 hours mean to you? Proceedings of the 5th USENIX Conference on File andStorageTechnologies(FAST'07)BestPaperAward.SanJose,CA,2007,1-9.
[33] BiancaSchroeder,Garth A.Gibson.Understanding FailuresinPetascaleComputers. JournalofPhysics:ConferenceSeries,78.2007,1-6.
[34] QiongLi,EnqiangZhou,ZhongLiu,YufengGuo,XuejunYang. High-PerformanceStoragefor DSM System Using Global Cache, the 10th IEEE International Conference on HighPerformanceComputing andCommunications,2008,971-975.
[35] DrorG. Feitelson, Peter F. Corbett, SandraJohnson Baylor, etal. ParallelI/O Subsystems inMassivelyParallelSupercomputers.IEEEParallel&distributedtechnology,1995,1-6.
[36] R. Jain, J. Werth, J. C. Browne. Input/output in parallel and distributed computersystemsKluwerAcademicPublishers,1996.
[37] JohnM.May.ParallelI/OforHighPerformanceComputingAcademicpress,2001.
[38] M. R. Fahey, J. Larkin. Guidelines for Efficient Parallel I/O on the Cray XT3/XT4.Proceedingsofthe49thCrayUserGroup(CUG07).2007,1-5.
[39] Weikuan Yu, Vetter, J. S, et al. Performance characterization and optimization of parallelI/O on the CrayXT. IEEE International Symposiumon Parallel and Distributed Processing,IPDPS2008. 2008,1-11.
[40] Rob Ross(ANL), Jose Moreira(IBM), Kim Cupps(LLNL), et al. Parallel I/O on the IBMBlueGene/LSystem.2007,1-5.
[41] Weikuan Yu, Sarp Oral, Jeffrey Vetter, et al. Efficiency Evaluation of Cray XT Parallel IOStack.2007,1-6.
[42] S.R. Alam, R. F. Barrett, M. R. Fahey, et al. Cray XT4: An Early Evaluation for PetaScaleScientific Simulation. in ACM/IEEE conference on High Performance Networking andComputing(SC07).Reno,NV,2007,1-9.
[43] M. Fahey, J. Larkin, J. Adams. I/O Performance on a Massively Parallel Cray XT3/XT4. in22nd IEEE International Parallel & Distributed Processing Symposium (IPDPS'08). Miami,FL,2008,1-8.
[44]庞征斌,李琼,李永进,张峻,徐炜遐,CC-NUMA系统分布共享I/O的数据一致性维护,计算机研究与发展,2007,44(suppl.):226-232.
[45] Shane Canon, H. Sarp Oral. A Center-Wide File System using Lustre. Oak Ridge NationalLaboratory,2006,1-4.
[46] James Laudon, Daniel Lenoski. The SGI Origin: a ccNUMA highly scalable server.Proceedings of the 24th Annual International Symposium on Computer Architecture(ISCA-97).NewYork,NY,USA,1997,241-251.
[47] SGI-Products:ServersandSupercomputers:SGIAltixFamily.http://www.sgi.com/products/servers/altix/,2008
[48] Windsor W. Hsu, Alan Jay Smith. The performance impact of I/O optimizations and diskimprovements.IBMJournalofresearchanddevelopment,2004,48(2):1-7.
[49] XiaoningDing;SongJiang;FengChen. A Buffer Cache Management Scheme ExploitingBothTemporalandSpatialLocalities[J].ACM TransactionsonStorage.2007,1-20.
[50] DavidLeong.CollaborativeObjectCachingforHeterogeneousOSDClusters.Proceedingsofthe2007ComputerScienceandITEducationConference.2007,426-436.
[51] M. Korupolu, M. Dahlin. Coordinated placement and replacement for large-scale distributedcaches.IEEETransactionsonKnowledgeandDataEngineering,2002,1317-1329.
[52] Wei-keng Liao, K. Coloma, A. Choudhary, et al. Cooperative Client-Side File Caching forMPI Applications[J]. International Journal of High Performance Computing Applications2007.2007,144-154.
[53] SongJiang, KeiDavis, Xiaodong Zhang.Coordinated MultilevelBufferCache Managementwith Consistent Access Locality Quantification. IEEE transactions on computers,2007,56(1):1-6.
[54] He Xubin, Ben Ou Li. A unified multiple-level cache for high performance storage systems.InternationalJournalofHighPerformanceComputingandNetworking2007. 2007,1-4.
[55] Guang R. Gao Brian Lucas. A High Throughput Massive I/O Storage Hierarchy forPETA-scale High-end Architectures. HEC FSIO Research and DevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-9.
[56] Jiang Song, Zhang Xiaodong. Coordinating Existing Prefetching Strategies for Multi-levelStorage Systems. HEC FSIO Research and Development Conference/HECURA FSIO PIMeeting'08.VA.Stafford,USA,2008,1-8.
[57] Z. Chen, Y. Zhang, H. Zhou, Y., Scott, et al. Empirical evaluation of multi-level buffercache collaboration for storage systems. ACM SIGMETRICS: Performance EvaluationReview,2005,33(1):1-7.
[58] Florin Isaila, Guido Malpohl. Integrating Collective I/O and Cooperative Caching into the"Clusterfile"ParallelFileSystem.ICS’04.Malo,France,2004,1-4.
[59] Yong Chen. Hiding I/O Latency with Pre-execution Prefetching for Parallel Applications.SC2008.Austin,Texas,2008,1-2.
[60] Surendra Byna. Parallel I/O Prefetching Using MPI File Caching and I/O Signatures.SC2008.Austin,Texas,2008,1-4.
[61] C. K. Yang, T. Mitra, T. Chiueh. A Decoupled Architecture for Application-Specific FilePrefetching.inFreenixTrackofUSENIX2002AnnualConference.2002,1-4.
[62] S. Jiang, X. Zhang. STEP: Sequentiality and Thrashing Detection Based Prefetching toImprove Performance of Networked Storage Servers. in 27th International Conference onDistributedComputingSystems(ICDCS'07).2007,1-4.
[63] YongChen, Surendra Byna, Xian-He Sun. DataAccessHistoryCache andAssociated DataPrefetchingMechanisms.SC07.Reno,Nevada,USA,2007,1-5.
[64] Xiaosong Ma.Coordinating ExistingPrefetching StrategiesforMulti-levelStorage Systems.HEC FSIO Research and Development Conference/HECURA FSIO PI Meeting '08. VA.Stafford,USA,2008,1-6.
[65] XiaosongMa.Application-adaptiveI/OStackforData-intensiveScientificComputing.HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '07. VA.Stafford,USA,2007,1-4.
[66] AveryChing Wei-keng Liao. Using MPI File Caching to Improve Parallel Write PerformanceforLarge-ScaleScientificApplications.SC07.Reno,Nevada,USA,2007,1-6.
[67] Wei-keng Liao. Scalable I/O Middleware and File System Optimizations forHigh-Performance Computing. HEC FSIO Research and DevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-7.
[68] Mahmut Kandemir, Alok Choudhary. Compiler-Directed I/O Optimization. Proceedings oftheInternationalParallelandDistributedProcessingSymposium(IPDPS.02).2002,1-10.
[69] A. Acharya, M. Uysal, J. Saltz. Active disks: Programming model, algorithms andevaluation.InProc.oftheACMASPLOSConference.1998,81-91.
[70] K. Keeton, D. A. Patterson, J. M. Hellerstein. A case for intelligent disks (IDISKs). InSIGMODRecord,Vol.24,No.7,1998,42-52.
[71] E. Riedel, G. Gibson, C. Faloutsos . Active storage for large-scale data mining andmultimedia.InProc.ofthe24thInt.Conf.VeryLargeDataBases(VLDB).1998,62-73.
[72] P. M. Kogge, J. B. Brockman. Processing in memory: Chips to petaflops. In Workshop onMixingLogicandDRAM:ChipsthatComputeandRemember(atISCA'97).1997,1-6.
[73] Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau, Lakshmi N. Bairavasundaram.Semantically-smart disk systems: Past, present, and future. ACM SIGMETRICSPerformance Evaluation Review on Design, implementation, and performance of storagesystems, 2006,33(4):29-35.
[74] Ma Xiaonan, Reddy A. L. N. MVSS.An active storage architecture. IEEE Trans on ParallelandDistributedSystems,2003, 14(10):993-1005.
[75] G. A. Gibson, D. F. Nagle, K. Amiri, et al.File server scaling with network-attached securedisks. In Proc. Of the 1997 ACM SIGMETRICS International Conference on MeasurementandModelingofComputerSystems.1997,272-284.
[76] M. Mesnier, G. Ganger, E. Riedel. Object-based storage. IEEE Communications Magazine,2005,.41(8):84-90.
[77] Azagury A, Dreizin v, Factor M. Towards an object store. In: Proc of the 20th IEEE/11thNASAGoddardConferenceonMassStorageSystemsandTechnologies(MSST2003).2003,165-176.
[78] Sivathanu M,Arpaci-Dusseau A.C,Arpaci-Dusseau R.H.Evolving RPC foractive storage.ACM SIGPLAN Notices SESSION on Communication abstractions and optimizations,2002, 37(10):264-276.
[79] H. Lim, V. Kapoor, C. Wighe, et al. Active disk file system: A distributed, scalable filesystem. In Proc. of the 18th IEEE Symposium on Mass Storage Systems and Technologies.SanDiego,2001,101-115.
[80] Joel C. Wu, Scott A. Brandt. The design and implementation of AQuA: an adaptive qualityof service aware object-based storage device. In Proceedings of the 23rd IEEE 14th NASAGoddard Conference on Mass Storage Systems and Technologies. College Park, MD, 2006,209-218.
[81] The advantages of object-based storage -secure, scalable, dynamic storage devices.http://www.Seagate.com,2008
[82] D. Nagle, D. Serenyi, A. Matthews. The panasas activescale storage cluster - Deliveringscalable high bandwidth storage. Proceedings of the 2004 AGM/IEEE Conference onSupercomputing(SC2004).2004,-10.
[83] O. Rodeh, A. Teperman. zFS - A Scalable distributed File System using Object Disks. InProceedings of the IEEE Mass Storage Systems and Technologies Conference. San Diego,CA,USA,2003,207-218.
[84] Feng Wang, Scott A. Brandt, Ethan L. Miller, et al. OBFS: A file system for object-basedstorage devices. In Proceedings of the 21st IEEE / 12th NASA Goddard Conference onMassStorageSystemsandTechnologies.CollegePark,MD,2004,283-300.
[85] Scott A. Brandt Sage A.Weil, Ethan L. Miller, Darrell D. E. Long, et al. Ceph: A scalable,high-performance distributed file system. In Proceedings of the 7th Symposium onOperatingSystemsDesignandImplementation(OSDI).Seattle,WA,2006,1-6.
[86] Sage A. Weil, Scott A. Brandt, Ethan L. Miller, et al. CRUSH: Controlled, scalable,decentralized placement of replicated data. In Proceedings of the 2006 ACM/IEEEConferenceonSupercomputing(SC'06).Tampa,FL,2006,1-7.
[87] PeteWyckoff. ApplicabilityofObject-Based Storage DevicesinParallelFile Systems. HECFSIO Research and Development Conference/HECURA FSIO PI Meeting '07. VA.Stafford,USA,2007,1-9.
[88] Brent Welch, Garth Gibson. Managing scalability in object storage systems for HPC Linuxclusters. In Proceedings of the 21st IEEE / 12th NASA Goddard Conference on MassStorageSystemsandTechnologies.2004,433-445.
[89] P.Braam.Thelustrestoragearchitecture.ClusterFileSystem,Inc,http://www.csis.hku.hk/cluster2003/presentation/vt/lustre.pdf,2007.
[90] E. J. Felix, K. Fox, K. Regimbal, etal.Active storage processing in a parallelfile system. InProc. of the 6th LCI International Conference on Linux Clusters: The HPC Revolution.2006,1-7.
[91] Juan Piernas. Evaluation of Active Storage Strategies for the LustreParallel File System. SC'07.Reno,Nevada,USA,2007,1-10.
[92] MitchellT.M.MachineLearning.McGraw-HillScience/Engineering/Math,1997,1-7.
[93] Sage Weil, Andrew Leung, Scott A. Brandt, et al. RADOS: A Fast, Scalable, and ReliableStorage Service for Petabyte-scale Storage Clusters. Proceedings of the ACM PetascaleDataStorageWorkshop2007(PDSW07).2007,1-5.
[94] Milo Polte, Jiri Simsa, Wittawat Tantisiriroj, et al. Fast Log-based Concurrent Writing ofCheckpoints. Proceedings of the 3rd Petascale Data Storage Workshop held in conjunctionwithSupercomputing'08.Austin,TX,2008,1-13.
[95] Eduardo Pinheiro, Wolf-Dietrich Weber, Luiz Andre Barros.Failure Trends in a Large DiskDrivePopulation.InFAST'07.SanJose,CA,2007,1-9.
[96] JonG.Elerath, MichaelPecht. Enhanced ReliabilityModeling ofRAID Storage Systems. InDSN'07.Edinburgh,UK,2007,1-7.
[97] Remzi Arpaci-Dusseau. Formal Failure Analysis for Storage Systems. HEC FSIO Researchand Development Conference/HECURA FSIO PI Meeting '07. VA. Stafford,USA, 2007,1-5.
[98] Qin Xin, Ethan L. Miller, Thomas J. E. Schwarz. Evaluation of distributed recovery inlarge-scale storage systems. In Proceedings of the 13th IEEE International Symposium onHighPerformanceDistributedComputing(HPDC).Honolulu,HI,2004,172-181.
[99] Soila Pertet, Rajeev Gandhi, Priya Narasimhan. Fingerpointing Correlated Failures inReplicated Systems. USENIX Workshop on Tackling Computer Systems Problems withMachineLearningTechniques(SysML).Cambridge,MA,2007,1-8.
[100] Muthian Sivathanu, Vijayan Prabhakaran, Andrea C. Arpaci-Dusseau, et al. ImprovingStorageSystemAvailabilitywithD-GRAID.InFAST'04.SanFrancisco,CA,2004,1-6.
[101] Lei Tian, Dan Feng, Hong Jiang, et al. PRO: A Popularity-based Multi-threadedReconstruction Optimization for RAID-Structured Storage Systems. In FAST '07. San Jose,CA,2007,1-9.
[102] Kiron Vijayasankar, Gopalan Sivathanu, Swaminathan Sundararaman. Exploiting Type-AwarenessinaSelf-RecoveringDisk.InStorageSS'07.Alexandria,Virginia,2007,1-8.
[103] Kevin Greenan, Ethan Miller, Thomas Schwarz, et al. Disaster Recovery Codes: IncreasingReliability with Large-Stripe Error Correction Codes. In StorageSS '07. Alexandria,Vir-ginia,2007,1-11.
[104] Haryadi S. Gunawi, Vijayan Prabhakaran, Swetha Krishnan, et al. Improving File SystemReliabilitywithI/OShepherding.InSOSP'07.Stevenson,WA,2007,1-7.
[105] http://msdn.microsoft.com/en-us/magazine/cc163388.aspx,2009
[106] http://tservice.net.ru/~s0mbre/old/?section=projects&item=pohmelfs, 2009
[107] http://commons.apache.org/transaction/file/index.html,2009
[108] B. Hong F. Wang, Q. Xin, S. A. Brandt, E. L. Miller, et al. File System Workload AnalysisFor Large Scale Scientific Computing Applications. in Proceedings of the 21st IEEE/12thNASAGoddardConferenceonMassStorageSystemsandTechnologies.CollegePark,MD,2004,1-3.
[109]卢凯.面向并行科学计算应用的并行文件系统技术研究,博士论文,国防科学技术大学,1999.
[110] DanielA.Reed. Workshop on the Road Map forthe Revitalization ofHigh End Computing.http://www.cra.org/reports/supercomputing.pdf,2003.
[111] Nitin Agrawal, Vijayan Prabhakaran, Ted Wobber, et al. Design tradeoffs for SSDperformance.InUSENIXAnnualTechnicalConference.Boston,MA,2008,1-4.
[112] Dushyanth Narayanan, Eno Thereska, Austin Donnelly. Migrating enterprise storage toSSDs:analysisoftradeoffs. TechnicalMicrosoftResearchCambridge,UK,2008.
[113] DataStorageTechnologies-Transitiontosolidstate.http://www.ru.is/kennarar/andri/nyti/papers2007/DST.pdf,2007.
[114] Kang J, Jo H, Kim J. A superblock-based flash translation for NAND flash memory. NewYork,NY,USA,2006,1-5.
[115] B. Banister, J. Bowen. High Performance QFS with Solid State Disk Metadata Storage.TechnicalTexasMemorySystems,Inc,2005.
[116] DavidDriggers.TheTop5TrendsinHighPerformanceComputingfor2009.http://www.wwpi.com/systems/systems/6663-the-top-5-trends-in-high-performance-computing-for-2009,2009.
[117] Jing Li, Patrick Ndai, Ashish Goel, et al. An Alternate Design Paradigm for RobustSpin-Torque Transfer Magnetic RAM (STT MRAM) from Circuit/Architecture Perspective.14thAsiaandSouthPacificDesignAutomationConference.2009,1-6.
[118] Kato, Yoshihisa Tanaka, Hiroyuki Isogai, et al. Embedded FeRAMChallengesin the 65-nmTechnology Node and Beyond. Applications of ferroelectrics, 2006. isaf '06. 15th ieeeinternationalsymposium.2006,1-4.
[119] Jin Hyuk Yoon, Eyec Hyun Nam, Yoon Jae Scong, et al. Chameleon: A High PerformanceFlash/FRAM Hybrid Solid State Disk Architecture. Computer Architecture Letters,20071-7.
[120] T. Rueckes, K. Kim, E. Joselevich. Carbon Nanotube-based Nonvolatile Random AccessMemoryForMolecularComputing.Science,2000,298:94-97.
[121] Mustafa Uysal, Arif Merchant, Guillermo A. Alvarez. Using MEMS-based storage in diskarrays. Proceedings of the 2nd USENIX Conference on File and Storage Technologies. SanFrancisco,CA,2003,1-9.
[122] R. F. FREITAS, W. W. WILCKE. Storage-class memory. The next storage systemtechnology.IBMjournalofresearchanddevelopment,2008, 52:1-5.
[123] G. W. Burr, B. N. Kurdi, J. C. Scott, et al. Overview of candidate device technologies forstorage-class memory. (IBM Journal of Research and development) Storage TechnologiesandSystems,2008, 52:4-5.
[124] Winfried Wilcke. Flash and Storage Class Memories: Technology Overview & SystemsImpact.HECFSIOResearchandDevelopmentConference/HECURAFSIOPIMeeting'08.VA.Stafford,USA,2008,1-8.
[125] An-I A. Wang, Peter Reiher, Gerald J. Popek, et al. Conquest: better performance through adisk/persistent-RAM hybrid file system. In Proceedings of the 2002 USENIX AnnualTechnicalConference.Monterey,2002,15-28.
[126] Ethan L. Miller. HeRMES: High-Performance Reliable MRAM-Enabled Storage. 8th IEEEWorkshop on Hot Topics in Operating Systems (HOTOS-VIII). Schloss Elmau, Germany,2001,83-87.
[127] I. Ari, M. Gottwals, D. Henze. SANBoost: Automated SAN-Level Caching in Storage AreaNetworks. In Proceedings of the 13th IEEE International Conference on AutonomicComputing(ICAC).2004,164-171.
[128] M.Casey.SANCache:SSDintheSAN.TechnicalSoliddata,Inc,2000.
[129] Gurumurthi S, Sivasubramaniam A, Kandemir M. DRPM: Dynamic Speed Control forPower Management in Server Class Disks. In Proc. of the 30th annual internationalsymposiumonComputerarchitecture.SanDiego,California,2003,169-181.
[130] Gurumurthi S, Zhang J, Sivasubramaniam A. Interplay of Energy and Performance for DiskArrays Running Transaction Processing Workloads. In Proc. of 2003 IEEE InternationalSymposium on Performance Analysis of Systems and Software. Austin, Texas; USA,2003,123-132.
[131] Adam Roth. Energy Conservation for Real-Time Disk Systems with I/O Burstiness. IEEEInt'lWorkshopNextGenerationAutonomousStorageandHighPerformanceComputing.St.Thomas,VirginIslands,2008,1-4.
[132] Mais Nijim. An Adaptive Energy-Conserving Strategy for Parallel Disk Systems. in theProceedings of the 12th IEEE International Symposium on Distributed Simulation and RealTimeApplications(DS-RT'08).Vancouver,BritishColumbia,Canada,2008,1-5.
[133] CharlesWeddle.PARAID:AGear-ShiftingPower-AwareRAID,1-7.
[134] Eduardo Pinheiro. Energy Conservation Techniques for Disk Array-Based Servers. ICS'04.Malo,France,2004,1-8.
[135] Seung Woo Son. EnergyAware Data Prefetching for MultiSpeed Disks. CF'06. Ischia, Italy,2006,1-3.
[136] LuisUseche.EXCES:EXternalCachinginEnergySavingStorageSystems,1-6.
[137] JianhuiYue.AnEnergy-Oriented Evaluation ofBufferCacheAlgorithmsUsingParallelI/OWorkloads. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2008, 19(11):1-9.
[138] DavidE.Culler,JaswinderPalSingh,AnoopGupta.并行计算机体系结构:硬件与软件结合的设计与分析,机械工业出版社,2002.
[139] Mark D. Hill, Anne E. Condon, Manoj Plakal, et al. A System-Level SpecificationFramework for I/O Architectures. the 11th Annual Symposium on Parallel Algorithms andArchitectures(SPAA).1999,1-6.
[140] John Wilkes, Richard Golding, Carl Staelin, et al. The HP AutoRAID heirarchical storagesystem.IEEETransactiononcomputersystems, 1996, 14(1):108-136.
[141] E. Anderson, M. Kallahalla, S. Spence, et al. Quickly finding near-optimal storage systemdesigns.http://hpl.hp.com/research/ssp/papers/ergastulum-paper.pdf,2005.
[142] Mohamed N. Bennani, Daniel A. Menasce. Resource allocation for autonomic data centersusing analytic performance models. In The 2nd IEEE International Conference onAutonomicComputing(ICAC-05).Seattle,WA,2005,229-240.
[143] Hewlett-PackardCompany.Adaptiveenterprise.http://www.hp.com/products1/promos/adaptive/enterprise/us/adaptive/enterprise.html,2005.
[144] M. Wang. Black-Box Storage Device Modeling with Learning:[PhD dissertation]. CarnegieMellonUniversity,2006.
[145] Engin Ipek, Onur Mutlu, Jose F. Martnez, et al. Self-Optimizing Memory Controllers. AReinforcement Learning Approach, Intl. Symp. on Computer Architecture (ISCA). Beijing,China,2008,1-8.
[146] Yu Zhang, Bharat Bhargava. Self-Learning Disk Scheduling. IEEE Transactions onknowledgeanddataengineering,2009,21(1):1-9.
[147] D. L. Martens, M. J. Katchabaw. Optimizing System Performance through Dynamic DiskScheduling Algorithm Selection,WSEAS Trans. Information Science and Applications,2006,1-7.
[148] M. Stillger, G. Lohman, V. Markl, et al. LEO-DB2'S Learning Optimizer. Proc. 27thinternationalconferenceof VeryLargeDataBases(VLDB).2001,1-5.
[149] K. Shen, M. Zhong, C. Li. I/O System Performance Debugging Using Model-DrivenAnomaly Characterization. Proc. Fourth Usenix Conf. File and Storage Technologies(FAST'05).2005,1-7.
[150] J. Wildstrom, P. Stone, E. Witchel, et al. Machine Learning for On-Line HardwareReconfiguration.Proc.20thInt'lJointConf.ArtificialIntelligence(IJCAI'07).2007,1-9.
[151] http://www.cs.waikato.ac.nz/ml/weka/,2008.
[152] William W. Cohen. Fast effective rule induction. In Proceedings of the 12th InternationalConferenceonMachineLearning(ICML-95).1995,15-123.
[153]舒继武,薛巍,付长冬.网络存储系统与技术的现状与发展趋势.中国计算机学会通信,2004,1-5.
[154] Maurice Herlihy, J. EliotMoss. TransactionalMemory:ArchitecturalSupportfor Lock-FreeData Structures. Proceedings of the 20th Annual International Symposium on ComputerArchitecture.1993,289--300.
[155] C. Scott Ananian, Krste Asanovic, Bradley C. Kuszmaul, et al. Unbounded TransactionalMemory. Proceedings of the Eleventh International Symposium on High-PerformanceComputerArchitecture.2005,316--327.
[156] Lance Hammond, Vicky Wong, Mike Chen, et al. Transactional Memory Coherence andConsistency. Proceedings of the 31st Annual International Symposium on ComputerArchitecture.2004,102.
[157] Kevin E. Moore, Jayaram Bobba, Michelle J. Moravan, et al. LogTM: Log-basedTransactional Memory. Proceedings of the 12th International Symposium onHigh-PerformanceComputerArchitecture.2006,254--265.
[158] Luke Yen, Jayaram Bobba, Michael M. Marty, et al. LogTM-SE: Decoupling HardwareTransactional Memory from Caches. Proceedings of the 13th International Symposium onHigh-PerformanceComputerArchitecture(HPCA).2007,1-7.
[159] JayaramBobba,NeelamGoyal, Mark D. Hill, etal. TokenTM: EfficientExecution of LargeTransactions with Hardware Transactional Memory. Proceedings of the 35th InternationalSymposiumonComputerArchitecture.IEEEComputerSociety,2008.2008,127-138.
[160] Maurice Herlihy, Victor Luchangco, Mark Moir, et al. Software Transactional Memory forDynamic-SizedDataStructures. 2003,92--101.
[161] DaveDice,OriShalev,NirShavit.TransactionalLockingII.2006,1-5.
[162] Bratin Saha, Ali- Reza Adl-Tabatabai, Richard L. Hudson, et al. McRT-STM: a HighPerformance Software Transactional Memory System for a Multi-core Runtime. Proc. 11thACM SIGPLAN Symp. on Principles and Practice of Parallel Programming (PPoPP '06).2006,187-197.
[163] Arrvindh Shriraman, Sandhya Dwarkadas, Michael L. Scott. Flexible DecoupledTransactional Memory Support. Proceedings of the 35th Annual International SymposiumonComputerArchitecture.2008,1-6.
[164] Cao Minh, Martin Trautmann, Jae Woong Chung, et al. An Effective Hybrid TransactionalMemory System with Strong Isolation Guarantees. Proceedings of the 34th AnnualInternationalSymposiumonComputerArchitecture.2007,1-9.
[165] Dice Yossi Dave, Dan Nussbaum. Early Experience with a Commercial HardwareTransactional Memory Implementation. Proceedings of the 14th International ConferenceonArchitecturalSupportforProgrammingLanguagesandOperatingSystems.2009,1-11.
[166] http://www.tevero.no/products/tffs/index.htm.
[167] http://opensolaris.org/os/community/zfs.
[168] Zhihui Zhang, Kanad Ghose. hfs: A hybrid file system prototype for improving small fileandmetadataperformance.InProceedingsofEuroSys2007.2007,1-4.
[169] DataDirectNetwor.DDNProducts:S2A9550.http://www.ddn.com,.
[170] EMCCLARiiONCX4.http://www.emc.com,.
[171] http://www.netapp.com.
[172] B. Banister, J. Bowen. High Performance QFS with Solid State Disk Metadata Storage.TechnicalTexasMemorySystems,Inc,2005.
[173] BoHong,ScottABrandt,DarrellELong,EthanL.Miller,etal.UsingMEMS-BasedStorageinComputerSystems-DeviceModelingandManagement.2006,1-7.
[174] FengWang,BoHong, Scott A. Brandt, Darrell D. E. Long. Using MEMS-based Storage toBoostDiskPerformance.2005,1-9.
[175] Steven W. Schlosser, Jiri Schindler, Anastassia Ailamaki, et al. Exposing and exploitinginternalparallelisminMEMS-basedstorage2003.
[176] PaulMassiglia.TheRAIDbook,6thEDITION.
[177] Yashushi Saito, Svend Frflund, Alistair Veitch. FAB: Building distributed enterprise diskarraysfromcommoditycomponents.InProceedingsofthe11thInternationalConferenceonArchitectural Support for Programming Languages and Operating Systems (ASPLOS).2004,48-58.
[178] John L. Gustafson. Reevaluating Amdahl's Law. Communications of the ACM,1988,31(5):532-533.
[179] A. Moga, A. Gefflaut, M. Dubois. Hardware Versus Software Implementation of COMA.Proceedingsofthe1997InternationalConferenceParallelProcessing.1997,248-255.
[180] P. Keleher, S. Dwarkadas, A. L. Cox. TreadMarks: Distributed Shared Memoryon StandardWorkstations and Operating Systems. In Proceedings of the Winter 1994 USENIXConference.1994,115-131.
[181] OpenMP:Simple,Portable,ScalableSMPProgramming.http://www.openmp.org/drupal/ .
[182] Rajesh A. Bordawekar. Quantitative Characterization and Analysis of the I/O Behavior of aCommercial Distributed-Shared-Memory Machine. IEEE Trans. Parallel Distrib. Syst.2000,11(5):509--526.
[183] L.Lamport.HowtoMakeaMultiprocessorComputerThatCorrectlyExecutesMultiprocessPrograms.IEEETransactionsonComputers,1979,28(9):690-691.
[184] M. Dubois, C. Scheurich, F. A. Briggs. Memory Access Buffering in Multiprocessors. Proc.13thInt'lSymp.onComputerArchitecture.1986,434-442.
[185] K. Gharachorloo, D. Lenoski. Memory Consistency and Event Ordering in ScalableShared-Memory Multiprocessors. In Proceedings of the 17th annual InternationalSymposiumonComputerArchitecture.Seattle,Washington,1990,15-26.
[186] J. P. Singh L.Iftode, K. Li Scope Consistency. A Bridge Between Release Consistency andEntry Consistency. In Proceedings of the 8th Annual ACM Symposium on ParallelAlgorithmsandArchitectures.1996,1-9.
[187] B. N. Bershad, M. J. Zekauskas, W. A. Sawdon. The Midway Distributed Shared MemorySystem. InProceedingsofthe38thIEEEInt'lComputerConf(COMPCONSpring'93).1993,528-537.
[188] Z. Huang, C. Sun, M. Purvis. View-based Consistency for Distributed Shared Memory. InProceedings of the 4th World Multiconference on Systemics, Cybernetics and Informatics(SCI'2000).2000,1-7.
[189] Z. Huang, C. Sun, S. Cranefield. View-based Consistency and Its Implementation. InProceedings of the 1st IEEE/ACM International Symposium on Cluster Computing and theGrid.Brisbane,2001,74-81.
[190] Z. Huang, C. Sun, M. Purvis. View-Based Consistency and False Sharing Effect inDistributedSharedMemory.InOperatingSystemsReview,SIGOPS, 2001, 35(2):51-60.
[191] Kourosh Gharachorloo. Memory Consistency Models for Shared-Memory Multiprocessors.PhD.Dissertation,1995,1-47.
[192] Yang Xuejun, Dai Huadong. Operating System-centric Memory consistency model--threadconsistency model. The Fourth International Workshop on Advanced Parallel ProcessingTechnologies.2001,1-8.
[193]戴华东.并行分布操作系统共享存储管理与优化技术研究,博士论文,国防科学技术大学,2002.
[194]胡伟武.共享存储系统结构,高等教育出版社,2001.
[195] Scheurich C, Dubois M. Correct Memory Operation of Cached-Based Multiprocessors.Proceeding of the 14th Anuual International Symposium on Computer Architecture. 1987,1-6.
[196] John L. Hennessy, David A. Patterson.计算机体系结构-量化研究方法机械工业出版社,2002.
[197]吴俊敏,杨超,陈国良.存储模型仿真器的设计与实现.计算机研究与发展,2004,42(3):394-403.
[198]庞征斌.基于SMP的CC-NUMA类大规模系统中Cache一致性协议研究与实现,博士论文,国防科学技术大学,2007.
[199] David E. Culler, Jaswinder Pal Singh, Anoop Gupta. Parallel Computer Architecture: AHardware/SoftwareApproach(SecondEdition)MorganKaufmannPress,1996.
[200] NAS Parallel Benchmarks I/O Version 3.2. NASA Advanced Supercomputing (NAS)Division,2005.
[201] P. Scheuermann, G. Weikum, P. Zabback. Data partitioning and load balancing in paralleldisksystems.VLDBJournal:VeryLarge DataBases,1998,7(1):48-66.
[202] IORBenchmark.http://www.llnl.gov/asci/purple/benchmarks/limited/ior,.
[203]莫则尧,许林宝,张宝琳.二维等离子体模拟粒子云网格方法的并行计算与性能分析.计算物理,1998,16(5):496-504.
[204] R. K. Sahoo H. Yu, C. Howson. High Performance File I/O for the Blue Gene/LSupercomputer,2007,1-6.
[205] Shenze Chen, John A. Stankovic, James F. Kurose, et al. Performance Evaluation of TwoNewDiskSchedulingAlgorithmsforReal-TimeSystem,1990.
[206] IozoneFilesystemBenchmark.http://www.iozone.org/src/current/iozone3_263.tar,2008.
[207] Elizabeth J. O'Neil;Patrick E. O'Neil, Gerhard Weikum. The lru-k page replacementalgorithm for database disk buffering. In Proceedings of the 1993 ACM SIGMODinternationalconferenceonManagementofdata.NewYork,NY,USA,1993,297-306.
[208] Donghee Lee, Jongmoo Choi, Jong-Hun Kim, et al. On the existence of a spectrum ofpolicies that subsumes the least recently used (LRU) and least frequently used (LFU)policies. In Proceedings of the 1999 ACM SIGMETRICS International conference onMeasurementandModelingofComputerSystems.1999,134-143.
[209] Y. Smaragdakis, S. Kaplan, P. Wilson. EELRU: Simple and Effective Adaptive PageReplacement.InProceedingsofthe1999ACMSIGMETRICSConferenceonMeasurementandModeling ofComputerSystems.1999,122-133.
[210] Yuanyuan Zhou, James Philbin, Kai Li. The multi-queue replacement algorithm for secondlevel buffer caches. In Proceedings of the General Track: 2002 USENIX Annual TechnicalConference.Berkeley,CA,USA,2001,91-104.
[211] Song Jiang, Xiaodong Zhang. LIRS: An efficient low inter-reference recency setreplacement policy to improve buffer cache performance. In Proceedings of the ACMSIGMETRICS Conference on Measurement and Modeling of Computer Systems. 2002,1-12.
[212] Nimrod Megiddo, Dharmendra S. Modha. ARC: A self-tuning, low overhead replacementcache. In Proceedings of the 2nd USENIX Conference on File and Storage Technologies(FAST).2003,115-130.
[213] Pei Cao, Edward W. Felten, Anna R. Karlin, et al. Implementation and performance ofintegrated application-controlled file caching, prefetching, and disk scheduling. ACMTransactionsonComputerSystems, 1996,14(4):311-343.
[214] R. Hugo Patterson, Garth A. Gibson, Eka Ginting, et al. Informed prefetching and caching.In Proceedings of the fifteenth ACM symposium on Operating systems principles (SOSP).NewYork,NY,USA,1995,79-95.
[215]TheodoreJohnson,DennisShasha.2Q:A lowoverhead highperformance buffermanagementreplacement algorithm. In Proceedings of the 20th International Conference on Very LargeDataBases(VLDB).SanFrancisco,CA,USA,1994,439-450.
[216] Gideon Glass, PeiCao. Adaptive page replacement based on memoryreference behavior. InProceedings of the ACM SIGMETRICS international conference on Measurement andModelingofComputerSystems.1997,115-126.
[217] Chris Gniady, Ali Raza Butt, Y. Charlie Hu. Program-counter-based pattern classification inbuffer caching. In Proceedings of 6th Symposium on Operating System Design andImplementation(OSDI).2004,395-408.
[218] Jong Min Kim, Jongmoo Choi, Jesung Kim, et al. A low-overhead, high-performanceunified buffer management scheme that exploits sequential and looping references. In 4thSymposiumonOperatingSystemDesignandImplementation(OSDI).2000,1-6.
[219] Feng Zhou, Rob von Behren, Eric Brewer. AMP: Program context specific buffer caching.InProceedingsoftheUSENIXTechnicalConference.2005,1-8.
[220] Jongmoo Choi, Sam H. Noh, Sang Lyul Min, et al. An implementation study of adetection-based adaptive block replacement scheme. In Proceedings of the 1999 USENIXAnnualTechnicalConference.1999,239-252.
[221] Yifeng Zhu. Cluster-based storage systems with high scalability:[PhD Thesis]. UniversityofNebraska,2005.
[222] J. Choi, S. Cho, S. H. Noh, et al. Analytic Prediction of Buffer Hit Ratios. IEEElectronicsLetters,2000,36(1):10-11.
[223] R. H. Patterson, G. A. Gibson, E. Ginting, et al. Informed Prefetching and Caching. InProceedingsofthe15thSymposiumonOperatingSystemPrinciples.1995, 1-16.
[224] G. R. Ganger, B. L. Worthington, Y. N. Patt. The DiskSim simulation environment version2.0 reference manual. Technical Carnegie Mellon University / University of Michigan,1999.