高性能分布计算的数据管理与优化

在高性能计算程序对海量分布存储数据的操控中,有效的数据管理很重要。该文提出一个新的高性能分布计算的数据管理与优化系统,它包含一个元数据管理系统和存储系统,提供一个容易使用且能自动进行存储访问优化的平台。该平台采用的多存储资源体系结构能够满足性能和存储容量需求,并能自适应地利用当前的I/O优化方法。     

基于SDN的数据中心网络资源调度机制

随着大数据应用的不断丰富, 现在的数据中心通常部署着多种集群计算框架, 并由统一的集群资源管理器(如Mesos)进行管理. 目前的集群资源管理主要关注计算资源和存储资源, 较少的涉及网络资源. 但研究表明高效的网络资源管理对于优化作业性能十分重要. 本文提出了一种基于SDN(Software Defined Network)的数据中心网络资源调度机制, 该机制可以根据管理员预设的网络资源分配策略, 加权的进行网络资源调度, 为高优先级的作业分配更多网络资源以优化性能, 并且实现不同作业之间的网络性能隔离. 我们基于开源SDN控制器实现了原型系统, 并通过实验验证了该机制的有效性.     

高强度I/O的应用对并行存储系统的挑战和解决方法研究

具有高I/O密集特性的高性能计算应用对高性能计算机存储系统综合性能的要求越来越高.以石油地震勘探数据处理为代表的一类重要应用表现出I/O数据量巨大、I/O访问密度大,对单个磁盘阵列存储部件的读写带宽要求高的特征.在Lustre文件系统中,充当对象存储服务功能的磁盘阵列设备输出带宽的不足将成为阻碍存储系统整体性能发挥的重要因素.针对此类问题,提出了一种缓存管理方法,分别在客户端添加VDISK模块,在OST端添加Cache模块,二者协同提高并行文件系统I/O的输出带宽的使用效率;另外,充分利用客户端空闲内存以及客户端之间的通信带宽,降低应用程序对磁盘阵列设备输出带宽的要求.通过大规模并行模型的验证表明,VDISK提高了实际可用的输出带宽,提高了外部存储系统的I/O效率.     

面向云存储的I/O资源效用优化调度算法研究

随着云计算的普及,越来越多的客户选择使用基于云的服务,以避免冗余的设施购买费用和繁杂的系统设计与维护,从而将精力集中在自己的专业领域.通常,云服务的客户从云服务供应商购买虚拟机,并根据双方商定达成的服务水平目标(service level objective,SLO)约束购买到的计算资源.分布式存储中大量的文件分布在不同的存储节点上,现有的CPU、内存以及带宽等资源的分配调度算法并不适用磁盘I/O资源.从云服务提供商的角度来说,高效用的I/O资源调度算法有利于提高其系统的利用率,节约资源开销并增加企业收益率.从云存储提供商为获取高效率高收益率的角度考虑,通过对用户的虚拟机在不同存储节点上的访问特性建模,提出了一个新的自适应分布式I/O资源调度算法,简称为PC算法.PC算法能够:1)根据用户与服务商之间制定的SLO,动态地在各个存储节点中为每个虚拟机制定适当的局部SLO,满足虚拟机对个体节点的访问需求;2)为各虚拟机提供高效健壮的资源分配策略,既能尽可能利用I/O资源,又避免由无序的I/O资源竞争导致的虚拟机I/O资源饥饿.PC算法能够根据不同的I/O资源供应状况在两种调度策略间自动切换,当系统I/O资源充足时,算法采用最早截止时间优先算法(earliest deadline first,EDF)方式提高I/O资源使用率;反之则根据每个I/O请求的预计效益来提高总收益率.实验结果表明,在不采用预先设定虚拟机对各个节点访问量的前提下,PC算法能根据访问模式制定合理的资源分配,提高系统的I/O资源利用率和收益.     

支持异构集群并行的高能物理数据处理系统

传统集群计算系统无法充分利用本地磁盘的存储资源和I/O,大量网络I/O成为系统瓶颈,导致资源利用率降低,并造成高昂的存储和网络成本.使用Hadoop处理分析作业可有效利用本地磁盘存储和I/O资源,而集群资源统一管理工具Mesos则使用轻量化的设计和高效的通信机制,能在不同计算集群之间动态共享集群资源.为此,分析高能物理数据处理的特点,利用Mesos构建异构集群间资源共享的高能物理实验数据处理系统,实现Torque/Maui和Hadoop集群的集成.测试结果表明,该系统能够在集群间动态分配集群资源,并利用本地存储和磁盘I/O显著降低网络I/O,提高集群资源利用率.     

云计算中存储资源管理策略的探究

本文将以Amazon云计算平台为例,介绍云计算中的存储资源管理策略,并在基于Xen的开源虚拟化平台上,使用I/O Controller等带宽分配管理工具,来模拟AWS S3等云存储系统中基于用户权重的分等级的资源分配策略。     

一种基于定制高性能互连的对象存储系统

基于高性能互连实现对象存储系统已经成为构建高性能计算机可扩展I/O系统的发展趋势。我们设计并实现了一种定制的高带宽、低延迟的高性能互连芯片HSNI,它提供了很好的通信性能,可用于构建对象存储系统。本文对HSNI的硬件体系结构、软件结构及其通信机制进行了介绍,并基于HSNI构建了高性能的对象存储系统。性能测试结果表明,HSNI芯片带宽高、延迟低,非常适合构建大规模对象存储系统,该存储系统能够很好地发挥Lustre系统的性能,并具有很好的可扩展性,能够很好地满足面向高性能计算的I/O系统需求。     

OHPSS:一种基于对象的高性能存储系统

现有存储系统主要存在以下问题:信息存储规模有限、系统结构限制了高速应用范围以及系统实时性和可靠性不够.为了满足大容量、高效、可扩展和高服务质量的存储系统需求,提出了一种基于对象的高性能存储体系结构--OHPSS.其特点是充分利用对象属性并采用适应性服务,达到提高访问性能的目的;基于对象存储设备(OBSD)有处理器和操作系统的硬件支持,既可避免系统性能瓶颈,又能提供异步I/O方式;存储磁盘采用SRAID(similar RAID)结构,提高了磁盘的并行性和可扩展性.测试结果表明,与其他存储系统相比,该存储结构显著提高了小块数据的随机读/写性能,并且多线程 异步I/O的工作方式也使系统扩展性得到了显著提高.     

分布式计算环境中的协同分配任务调度仿真系统

协同分配是在分布式计算环境中进行资源分配的一种重要技术,用于把一个应用程序分解为多个子作业,然后将其分配到多个资源上同时处理来满足特定的性能要求。本文提出了一个离散事件驱动的网格资源协同分配仿真系统,实现了对用户、调度器、协同分配器、协同预留器等协同分配相关实体的仿真,实现了FCFS、FPFS、Backfill等主要的协同分配调度算法和策略,可用于资源协同分配相关的分布式计算环境的资源管理和调度算法的仿真和研究。     

基于OpenStack的Swift负载均衡算法

为了解决由于OpenStack的负载分发不均衡而引发的存储性能下降、资源利用率降低、I/O响应时长增加等问题,提出对加权最小连接调度算法进行改进. 通过对对象存储的负载均衡调度算法研究,利用存储节点的CPU、内存、硬盘、I/O资源利用率信息,并结合节点任务请求连接数,计算存储节点负载能力、性能和权值. 负载均衡器根据每个存储节点的权值大小判断任务分发方向. 经实验证明改进的负载均衡调度算法能够解决存储读写性能下降的问题,提升数据吞吐率、存储读写性能和系统稳定性.     

I/O-aware bandwidth allocation for petascale computing systems

In the Big Data era, the gap between the storage performance and an application’s I/O requirement is increasing. I/O congestion caused by concurrent storage accesses from multiple applications is inevitable and severely harms the performance. Conventional approaches either focus on optimizing an application’s access pattern individually or handle I/O requests on a low-level storage layer without any knowledge from the upper-level applications. In this paper, we present a novel I/O-aware bandwidth allocation framework to coordinate ongoing I/O requests on petascale computing systems. The motivation behind this innovation is that the resource management system has a holistic view of both the system state and jobs’ activities and can dynamically control the jobs’ status or allocate resource on the fly during their execution. We treat a job’s I/O requests as periodical sub-jobs within its lifecycle and transform the I/O congestion issue into a classical scheduling problem. Based on this model, we propose a bandwidth management mechanism as an extension to the existing scheduling system. We design several bandwidth allocation policies with different optimization objectives either on user-oriented metrics or system performance. We conduct extensive trace-based simulations using real job traces and I/O traces from a production IBM Blue Gene/Q system at Argonne National Laboratory. Experimental results demonstrate that our new design can improve job performance by more than 30%, as well as increasing system performance.     

Adaptive hierarchical scheduling policy for enterprise grid computing systems

In an enterprise grid computing environments, users have access to multiple resources that may be distributed geographically. Thus, resource allocation and scheduling is a fundamental issue in achieving high performance on enterprise grid computing. Most of current job scheduling systems for enterprise grid computing provide batch queuing support and focused solely on the allocation of processors to jobs. However, since I/O is also a critical resource for many jobs, the allocation of processor and I/O resources must be coordinated to allow the system to operate most effectively. To this end, we present a hierarchical scheduling policy paying special attention to I/O and service-demands of parallel jobs in homogeneous and heterogeneous systems with background workload. The performance of the proposed scheduling policy is studied under various system and workload parameters through simulation. We also compare performance of the proposed policy with a static space–time sharing policy. The results show that the proposed policy performs substantially better than the static space–time sharing policy.     

Client-side straggler-aware I/O scheduler for object-based parallel file systems

Object-based parallel file systems have emerged as promising storage solutions for high-performance computing (HPC) systems. Despite the fact that object storage provides a flexible interface, scheduling highly concurrent I/O requests that access a large number of objects still remains as a challenging problem, especially in the case when stragglers (storage servers that are significantly slower than others) exist in the system. An efficient I/O scheduler needs to avoid possible stragglers to achieve low latency and high throughput. In this paper, we introduce a log-assisted straggler-aware I/O scheduling to mitigate the impact of storage server stragglers. The contribution of this study is threefold. First, we introduce a client-side, log-assisted, straggler-aware I/O scheduler architecture to tackle the storage straggler issue in HPC systems. Second, we present three scheduling algorithms that can make efficient decision for scheduling I/Os while avoiding stragglers based on such an architecture. Third, we evaluate the proposed I/O scheduler using simulations, and the simulation results have confirmed the promise of the newly introduced straggler-aware I/O scheduler.     

Fine-grained management of I/O optimizations based on workload characteristics

With the advent of new computing paradigms, parallel file systems serve not only traditional scientific computing applications but also non-scientific computing applications, such as financial computing, business, and public administration. Parallel file systems provide storage services for multiple applications. As a result, various requirements need to be met. However, parallel file systems usually provide a unified storage solution, which cannot meet specific application needs. In this paper, an extended file handle scheme is proposed to deal with this problem. The original file handle is extended to record I/O optimization information, which allows file systems to specify optimizations for a file or directory based on workload characteristics. Therefore, fine-grained management of I/O optimizations can be achieved. On the basis of the extended file handle scheme, data prefetching and small file optimization mechanisms are proposed for parallel file systems. The experimental results show that the proposed approach improves the aggregate throughput of the overall system by up to 189.75%.     

Optimization of cloud computing task execution time and user QoS utility by improved particle swarm optimization

In order to optimize the quality of service (QoS) and execution time of task, a new resource scheduling based on improved particle swarm optimization (IPSO) is proposed to improve the efficiency and superiority. In cloud computing, the first principle of resource scheduling is to meet the needs of users, and the goal is to optimize the resource scheduling scheme and maximize the overall efficiency. This requires that the scheduling of cloud computing resources should be flexible, real-time and efficient. In this way, the mass resources of cloud computing can effectively meet the needs of the cloud users. Field Programmable Gate Arrays (FPGA), high performance and energy efficiency in one field. Most of them would have been the particle algorithm. The current technological development is still in-depth at super-resolution image research at an unprecedentedly fast pace. In particular, systemic origin applications get a lot of attention because they have a wide range of abnormal results. The scientific resource scheduling algorithm is the key to improve the efficiency of cloud computing resources distribution and the level of cloud services. In addition, the physical model of cloud computing resource scheduling is established. The performance of the IPSO algorithm applied to cloud computing resource scheduling is analysed in the design experiment. The comparison result shows that the new algorithm improves the PSO by taking full account of the user's Qu's requirements and the load balance of the cloud environment. In conclusion, the research on cloud computing resource scheduling based on IPSO can solve the problem of resource scheduling to a certain extent.     

一种基于磁盘介质的网络存储系统缓存

随着计算规模越来越大,网络存储系统应用领域越来越广泛,对网络存储系统I/O性能要求也越来越高.在存储系统高负载的情况下,采用低速介质在客户机和网络存储系统的I/O路径上作为数据缓存也变得具有实际的意义.设计并实现了一种基于磁盘介质的存储系统块一级的缓存原型D-Cache.采用两级结构对磁盘缓存进行管理,并提出了相应的基于块一级的两级缓存管理算法.该管理算法有效地解决了因磁盘介质响应速度慢而带来的磁盘缓存管理难题,并通过位图的使用消除了磁盘缓存写Miss时的Copy on Write开销.原型系统的测试结果表明,在存储服务器高负载的情况下,缓存系统能够有效地提高系统的整体性能.     

Courier: Multi-dimensional QoS guarantees for the consolidated storage system

The increasing cost and complexity of data management is driving data centers to consolidate resource and provide storage service for multiplex applications. Therefore, storage systems must be able to guarantee multi-dimensional Quality of Service (QoS) for various applications. However, satisfying performance targets for each workload is challenging, because that the I/O characteristics of workloads usually varies widely and capability of storage system changes significantly. In this paper, we design and implement a novel QoS scheduler, Courier, to maintain satisfactory performance for applications even in this highly-volatile scenario. Courier dynamically alternates between a feedback-based latency controller and reward budget-based scheduling to achieve per-application performance requirement. The feedback-based controller is employed to estimate request service times and adjust scheduling strategy dynamically. Based on the estimation, it can identify time-critical requests from throughput-sensitive requests and schedule applications with time-critical requests preferentially to avoid latency violations. In addition, Courier rewards well-behavior application with more budget to maintain high storage utilization while providing performance guarantees. We evaluate the effectiveness of our approach using synthetic and real workloads, and the results show that Courier has good ability to achieve per-application performance targets.     

负载自适应的存储池数据重构策略

大数据时代各应用领域对计算机存储系统的性能和可靠性需求与日俱增。新型存储介质为计算机存储系统的性能提升提供了良好的机遇,基于固态盘的存储阵列（RAIS）已在各种存储系统中广泛使用。传统RAIS系统中当一块固态盘出现故障时,通过数据重构操作恢复故障盘的数据,重构时间长,且影响对上层应用提供I/O访问服务的能力。针对该问题,设计实现了基于多线程并发处理的存储池架构,该架构能够并发处理存储池中的I/O请求,提高用户I/O和数据重构I/O的访问性能。提出了一种负载自适应的I/O调度策略,能够在保证用户I/O服务质量的同时,提升数据重构效率。实验结果表明,基于存储池的多线程并发I/O处理架构能够提升数据重构性能,负载自适应的I/O调度策略能够根据用户I/O的负载情况动态调整用户I/O和数据重构I/O的调度比例,在保证用户I/O服务质量的同时,提升数据重构效率。     

GCSS: a global collaborative scheduling strategy for wide-area high-performance computing

Wide-area high-performance computing is widely used for large-scale parallel computing applications owing to its high computing and storage resources. However, the geographical distribution of computing and storage resources makes efficient task distribution and data placement more challenging. To achieve a higher system performance, this study proposes a two-level global collaborative scheduling strategy for wide-area high-performance computing environments. The collaborative scheduling strategy integrates lightweight solution selection, redundant data placement and task stealing mechanisms, optimizing task distribution and data placement to achieve efficient computing in wide-area environments. The experimental results indicate that compared with the state-of-the-art collaborative scheduling algorithm HPS+, the proposed scheduling strategy reduces the makespan by 23.24%, improves computing and storage resource utilization by 8.28% and 21.73% respectively, and achieves similar global data migration costs.     

高性能计算系统中的服务质量研究

高性能计算系统的资源管理以集群作业管理为主,这种粗粒度的管理方式缺乏有效的作业资源控制手段,不能准确了解作业的资源需求,在一定程度上仍然不可避免计算资源的浪费.针对高性能计算系统中高效利用系统计算资源的问题,提出并实现了基于操作系统的QoS服务质量框架,对作业资源使用进行细粒度的统计与控制,实现了资源的动态控制与协商机制,完善作业加载与调度策略,在高效利用系统资源方面取得了较好的应用效果.