PVFS客户端目录缓存设计与实现

缓存技术是提高并行文件系统性能的关键性技术。在并行文件系统中实现客户端目录缓存，不仅可以减轻目录服务器的压力，避免目录服务器成为系统瓶颈，而且可以简化客户端操作过程，提高并行文件系统的性能。该文对PVFS并行文件系统进行了分析，建立了客户端目录缓存模型；就客户端目录缓存实现的一些关键性问题，如缓存池开辟位置、一致性等问题进行了研究，给出了解决方法，并在此基础上实现了一个客户端目录缓存的原型系统。测试结果表明，加入缓存后，PVFS系统性能有所提高。     

Grid VFS中代理管理客户端磁盘缓存的设计与实现

设计并实现了在网格虚拟文件系统Grid VFS客户端采用代理管理磁盘缓存来隐藏网络延迟，提高网格数据利用率。此方法扩展了虚拟化分布式文件系统，提高了对网格资源有效完整的数据访问。在Grid VFS中，客户端代理动态地产生和管理磁盘缓存，每一个未修改的应用程序和OS都能应用，支持write-back策略、文件系统之间缓存的共享以及磁盘缓存的多级化，通过用户级VFS可以完整的集成到网格应用程序和资源中，从而提高了网格数据的访问性能。     

一种基于对象存储系统的元数据缓存实现方法

对象存储系统中元数据访问速度是影响文件系统性能的关键因素之一。提出了一种在客户端实现元数据缓存的方法,并用元数据操作协议保证缓存一致性,基于Hash的LFU-DA算法提高缓存查找效率。实验表明该方法减少了系统平均服务响应时间,提高了系统的I／O性能。     

基于对象存储文件系统的协作式缓存一致性策略

基于对象存储文件系统引入了客户端缓存机制以加快文件访问速度,由此也带来了缓存数据一致性问题。针对多客户端访问共享文件时缓存更新策略存在的延时、网络和服务器负载过大及网络消耗大等问题,基于Lustre对象存储文件系统实现了一种改进的协作式缓存一致性策略。实验结果表明,改进后的缓存更新策略,不仅可以保证缓存数据的一致性,而且减少了客户端与服务器交互次数,降低了访问延迟,提高了系统I/O性能。     

高性能集群文件系统的研究

集群系统提供了强大的批处理和并行计算的能力,具有高性能、高可扩展性、高吞吐量和易用性等特点,但是I／O性能和处理器性能的不匹配使得I／O成为许多应用的瓶颈,特别是处理大量数据的应用就更是如此。针对集群系统当前的现状,克服该瓶颈的常用方法就是采用一种并行虚拟文件系统（PVFS）技术。随着Linux群集系统性能的持续提高,高速并行文件已成为并行计算的一个必备部分。并行虚拟文件系统（PVFS）为高性能计算（HPC）群集和大型I／O密集并行应用提供了这样一个文件系统。首先介绍了PVFS的结构;然后研究了PVFS的存取和管理机制;最后分析,PVFS的工作原理。     

分类的缓存数据延迟写技术

根据缓存数据在缓冲区的活动性不同而设计和实现了一个分类的延迟写（Write behind）技术,通过将不同活动性的数据分类缓存，并延迟刷新到磁盘来减少、合并写磁盘的次数，从而改进文件系统的写性能．初步的实现结果表明，分类的缓存延迟写技术比LRU的缓存策略有更短的系统响应时间，写文件的反馈时间减少了11．3％，并且使用RWB策略的缓存命中率比使用LRU策略高．     

一个提高机群文件系统吞吐率的方法及其实现

随着工作站机群系统的迅猛发展,I/O瓶颈问题也日益严重。性能优良的机群文件系统对机群系统而言至关重要。吞吐率是指在单位时间内所处理的作业数,它是评价系统性能的一个重要指标。论文提出用路径名的多成分查找来提高文件系统的创建删除吞吐率,给出并在PVFS上实现了一个路径名多成分查找方法。文章的最后对PVFS进行了优化前后的性能比较,并给出了自己的结论。     

Sloth：一种新的适于缓存服务器的应用层文件系统

针对传统的文件系统(如UFS等)在支持缓存服务器时存在着元数据一致性维护、同步写操作、内存拷贝和多缓存诸多固有的缺陷，我们设计和实现了一种新的、高效的、可移植性好的文件系统Sloth。该系统在应用层实现，采用异步写操作和聚集文件等技术。仿真实验表明，Sloth文件系统有效地提高了读写磁盘的性能，大大减少了访问磁盘的次
数。     

一种加速广域文件系统读写访问的缓存策略

针对广域网高延迟、低带宽的特性给广域文件系统访问带来的性能影响问题,提出了一种不依赖于底层文件系统、能够加速广域文件系统读写访问的缓存策略.该策略支持基于区间粒度的文件数据缓存及访问,并支持元数据本地缓存;该策略提供基于阈值的容量管理功能,采用超时与最终一致相结合的方式维护缓存的一致性.最后使用典型的文件I/O基准测试工具和元数据性能测试工具对该缓存策略进行了评测,实验结果表明:该缓存策略减少了客户端与服务器的交互次数,给广域文件系统的数据访问带来了明显的性能提升,当缓存命中时其数据的读写性能与本地文件系统相近.     

分布式存储系统中客户端的设计与实现

BlueOcean是基于对象存储技术的大规模分布式存储系统,详细描述了其客户端软件的设计。客户端基于用户态文件系统fuse框架进行开发,既保证了客户端的通用性,又降低了开发和维护的复杂度。客户端实现了常用的posix接口,可支持绝大多数应用程序的透明运行;设计了一套高效的缓存机制,减少了元数据访问过程中的通信开销,减小了读写延迟,有效地提高了BlueOcean存储系统的性能。     

一种高性价比的PVFS并行文件系统

分析PVFS并行文件系统的构成,得出客户机软件、元数据服务器软件和数据服务器软件之间的接口关系,然后研究一种由PC客户机、PC元数据服务器和低价数据服务器共同构成的PVFS系统,其中客户及与元数据服务器不做重要改变,数据服务器软件需要开发修改以适应新的硬件平台,使得以更低的成本实现相同的系统或者以相同的硬件成本实现更高的性能。     

Disconnected operation for heterogeneous servers: A practical approach

This paper describes an implementation of a client caching system which supports disconnected and partially connected modes of operation while preserving the traditional role of client caching (improved response time) in the fully connected mode. The system is designed to facilitate various server-dependent verification protocols. The need for the heterogeneous server support stems from the fact that the mobile user is likely to access various network services, transfer data between them and use the data in the local unit application programs. The unique feature of this project is its implementation which does not require modification of the native system's kernel source code. The paper shows that an installable file system is a viable mechanism for providing such file system extensions without source code modifications. The results of our preliminary experiments indicate efficiency of the design and the feasibility of the implementation mechanism used.     

Retrospective adaptive prefetching for interactive Web GIS applications

A major task of a Web GIS (Geographic Information Systems) system is to transfer map data to client applications over the Internet, which may be too costly. To improve this inefficient process, various solutions are available. Caching the responses of the requests on the client side is the most commonly implemented solution. However, this method may not be adequate by itself. Besides caching the responses, predicting the next possible requests from a client and updating the cache with responses for those requests together provide a remarkable performance improvement. This procedure is called “prefetching” and makes caching mechanisms more effective and efficient. This paper proposes an efficient prefetching algorithm called Retrospective Adaptive Prefetch (RAP), which is constructed over a heuristic method that considers the former actions of a given user. The algorithm reduces the user-perceived response time and improves user navigation efficiency. Additionally, it adjusts the cache size automatically, based on the memory size of the client’s machine. RAP is compared with four other prefetching algorithms. The experiments show that RAP provides better performance enhancements than the other methods.     

Extending Proxy Caching Capability: Issues and Performance

Proxy caching is an effective approach to reduce the response latency to client requests, web server load, and network traffic. Recently there has been a major shift in the usage of the Web. Emerging web applications require increasing amount of server-side processing. Current proxy protocols do not support caching and execution of web processing units. In this paper, we present a weblet environment, in which, processing units on web servers are implemented as weblets. These weblets can migrate from web servers to proxy servers to perform required computation and provide faster responses. Weblet engine is developed to provide the execution environment on proxy servers as well as web servers to facilitate uniform weblet execution. We have conducted thorough experimental studies to investigate the performance of the weblet approach. We modify the industrial standard e-commerce benchmark TPC-W to fit the weblet model and use its workload model for performance comparisons. The experimental results show that the weblet environment significantly improves system performance in terms of client response latency, web server throughput, and workload. Our prototype weblet system also demonstrates the feasibility of integrating weblet environment with current web/proxy infrastructure.     

PVFS文件系统吞吐率分析和改进

1.引言近年来工作站机群系统以其低成本、高性能而日益成为高性能计算的潮流。尽管机群系统为许多应用提供丁高性能的计算平台,但是其相对低速的文件系统成为了系统的瓶颈,因此提高机群文件系统的性能具有重要的意义。吞吐率是指在单位时间内所处理的作业数,它是评价系统性能的一个重要指标。PVFS是一个机群文件系统,目前在大规模并行计算     

A Framework for Cache Management for Mobile Databases: Design and Evaluation

In a mobile computing environment, database servers disseminate information to multiple mobile clients via wireless channels. Due to the low bandwidth and low reliability of wireless channels, it is important for a mobile client to cache its frequently accessed database items into its local storage. This improves performance of database queries and improves availability of database items for query processing during disconnection. In this paper, we investigate issues on caching granularity, coherence strategy, and replacement policy of caching mechanisms for a mobile environment utilizing point-to-point communication paradigm.We first illustrate that page-based caching is not suitable in the mobile context due to the lack of locality among database items. We propose three different levels of caching granularity: attribute caching, object caching, and hybrid caching, a hybrid approach of attribute and object caching. Next, we show that existing coherence strategies are inappropriate due to frequent disconnection in a mobile environment, and propose a cache coherence strategy, based on the update patterns of database items. Via a detail simulation model, we examine the performance of various levels of caching granularity with our cache coherence strategy. We observe, in general, that hybrid caching could achieve a better performance. Finally, we propose several cache replacement policies that can adapt to the access patterns of database items. For each given caching granularity, we discover that our replacement policies outperform conventional ones in most situations.     

一种基于副本的PVFS负载均衡机制

并行虚拟文件系统PVFS的数据服务器缺少负载均衡机制,因此存在热点服务器,降低了系统整体性能.提出了一种基于副本的负载均衡机制,通过文件数据备份的方式进行负载迁移,以解决这一瓶颈问题.其通过选择备份文件时权衡文件的热度与大小以降低数据备份的开销,将热点数据以较小代价转移到较空闲的服务器上,有效地提高了整个系统的数据吞吐量.其主要涉及了热点监测、数据备份源-目的节点选择以及备份文件策略3个部分的工作.实验结果表明:提出的负载均衡机制有效地提高了系统的整体性能,最高达到了24%.     

DESOLA: An active linear algebra library using delayed evaluation and runtime code generation

Active libraries can be defined as libraries which play an active part in the compilation, in particular, the optimisation of their client code. This paper explores the implementation of an active dense linear algebra library by delaying evaluation of expressions built using library calls, then generating code at runtime for the compositions that occur. The key optimisations in this context are loop fusion and array contraction.Our prototype C++ implementation, DESOLA, automatically fuses loops arising from different client calls, identifies unnecessary intermediate temporaries, and contracts temporary arrays to scalars. Performance is evaluated using a benchmark suite of linear solvers from ITL (Iterative Template Library), and is compared with MTL (Matrix Template Library), ATLAS (Automatically Tuned Linear Algebra) and IMKL (Intel Math Kernel Library). Excluding runtime compilation overheads (caching means they occur only on the first iteration), for larger matrix sizes, performance matches or exceeds MTL; when fusion of matrix operations occurs, performance exceeds that of ATLAS and IMKL.     

Web caching with consistent hashing

A key performance measure for the World Wide Web is the speed with which content is served to users. As traffic on the Web increases, users are faced with increasing delays and failures in data delivery. Web caching is one of the key strategies that has been explored to improve performance.An important issue in many caching systems is how to decide what is cached where at any given time. Solutions have included multicast queries and directory schemes.In this paper, we offer a new Web caching strategy based on consistent hashing. Consistent hashing provides an alternative to multicast and directory schemes, and has several other advantages in load balancing and fault tolerance. Its performance was analyzed theoretically in previous work; in this paper we describe the implementation of a consistent-hashing-based system and experiments that support our thesis that it can provide performance improvements.