面向网络性能优化的虚拟计算资源调度机制研究

针对基于Xen的vCPU调度机制对虚拟机网络性能的影响进行了深入研究和分析。提出一种高效、准确、轻量级的网络排队敏感类型虚拟机（NSVM）识别方法,可根据当前虚拟机I/O传输特征将容易受到影响的虚拟机进行准确识别和区分。进而设计一种新型虚拟计算资源调度和分配机制Diff-Scheduler,将不同类型虚拟机的vCPU实施分池隔离调度,同时提高NSVM类型虚拟机vCPU的调度频率。原型系统实验结果表明,相比Xen默认的调度机制,Diff-Scheduler能够大幅提高虚拟机网络性能,同时保证计算资源分配的公平性。     

基于Linux的高性能集群的构建和性能优化

提出并实现了一种并行计算集群搭建的方案,采用国际通用的HPL标准进行浮点运算性能的测试。在优化集群高性能运算性能的方面,对传统基于经验的穷举办法进行优化,提出一种基于参数影响因子排序的优化方案,有效地减少了穷举法盲目试验的时间。另外次算法还具有一定的开放性,以往的经验数据可以得到有效利用。     

Thermal design optimization of novel modular power converter assembly enabling higher performance,reliability and availability

An alternative integration scheme for a half-bridge switch using 70 μm thin Si IGBTs and diodes is presented. This flat switch, which is designed for high-frequency application with high power density, exhibits high strength, high toughness, low parasitic inductance and high thermal conductivity. Such a novel assembly approach is suitable to optimize performance, reliability and availability of the power system in which it is used. The paper focuses on the thermal performance of this assembly at normal and extreme operating conditions, studied by means of FEM thermo-fluidynamic simulations of the module integrated with connectors and liquid cooler, and thermal measurement performed on an early prototype. Improved solutions are also investigated by the FE model.     

Game based capacity allocation for utility computing environments

Utility computing has the potential to greatly increase the efficiency of IT operations by sharing resources across multiple users. This sharing, however, introduces complex problems with regards to pricing and allocating these resources in a way that is fair, easy to implement, and economically efficient. In this paper, we study a queue-based model that attempts to address these issues. Each client/user has a continuous flow of jobs that need to be processed. The service rate each receives, however, is proportional to a bid it submits to the system operator. Assuming that user costs are some function of their average backlogs plus their bid amounts, we use this allocation mechanism to construct an economic game. Much previous research has shown that these types of allocation games have desirable properties if the cost functions are well-defined and convex over the space of possible outcomes. Because of its queueing interface, however, our model induces functions that do not satisfy the latter, commonly assumed properties. In spite of these complications, we show that the game still has a unique equilibrium and that the system will converge to this point if users iteratively make ??best response?? updates to their bids. Finally, we explore the ??price of anarchy?? in our model, proving a bound on efficiency losses as a function of several fundamental system parameters. Thus, our scheme results in equilibria with a number of highly desirably properties.     

基于高性能运算PC集群综合优化技术研究

随着Pc集群技术在高性能运算地震资料处理领域的迅猛发展,如何提高PC集群的整体应用性能,成为业界普遍关注的技术难题。针对应用软件的具体需求,收集、整理集群节点的性能数据,定位影响作业运行效率的系统瓶颈,通过参数优化、硬件升级、系统定制和脚本开发等技术手段,全方位实现地震资料处理系统的最优化运行,从而达到提升PC集群系统整体性能的目的,同时也为高性能运算Pc集群的性能优化提供了可参考和借鉴的解决方案。     

远距离输砂在线效益优化自动控制系统

远距离疏浚工程需要安装多个加压泵以克服沿程输送阻力,多级加压泵的同时运行耗能较多,提高了系统运行成本.本文针对现有效益优化方法在实时控制方面的不足,设计实现了输砂在线效益优化控制系统.在分析产量和能耗与含砂量关系模型的基础上,建立了输砂效益优化目标函数,并根据黄金分割法设计了在线优化算法.实验结果显示,本文提出方法可将输砂效益提升20％,且具有较好的稳定性,能显著提升输砂系统运行效率.     

An approximate protocol analysis with performance optimization for WDM networks

In this study, our main goal is to investigate the performance optimization conditions for WDM networks. We introduce a network architecture of passive star topology that uses a Multi-channel Control Architecture (MCA) to avoid both the data channels and the receiver collisions. Especially, we propose a synchronous access scheme that exploits the propagation delay parameter in order to assign the data channels to the stations for successful data packet transmission. Thus, we achieve effective bandwidth utilization. An approximate analysis based on Poisson statistics is developed in order to explore the performance measures optimization. Finally, extensive comparative study is given for various stations populations and number of MCA channels.     

Study on utility optimization for randomized response mechanism

For the study of privacy-utility trade-off in local differential privacy,the utility optimization models of binary generalized random response mechanism for the case of differential privacy and approximate differential privacy were established.By graphic method,optimality proof,software solution and extreme point method,the exact expression of the optimal utility with privacy budget and the distribution of input data was obtained,and the corresponding optimal randomized response mechanism was given.The results show that both the optimal utility and optimal mechanism are related to privacy budget and input data distribution.Moreover,the discussion for multivariate randomized response mechanism shows that the method of extreme points of local differential privacy is feasible to the solution.     

An efficient procedure for checking performance limits in servodrive selection and optimization

Many mechatronic systems involve complex motions resulting in highly dynamic motor loads. Motor characteristics, such as rotor inertia, affect this motor load, which makes the design and selection of the complete drive system-consisting of motor, mechanical reducer, and converter-a huge combinatorial problem. The approach presented in the paper separates the problem into feasibility and optimization phases, and normalizes it with respect to the rotor inertia. This yields a fast and automated design procedure for servo drive systems, giving more insight through a set of conveniently arranged graphs     

Disk system architectures for high performance computing

Following a brief review of the fundamentals of disk system architecture, the characteristics of the applications that demand high I/O system performance are described. Conventional ways to improve disk performance are discussed. New developments in disk array systems are introduced, and controller architectures are described     

备件保障度评估与备件需求量模型研究

基于更新过程理论,给出了部件寿命服从不同分布(指数分布、威布尔分布、正态分布和Г分布)情况下,单不可修部件的备件保障度模型;然后给出了多不可修部件和多可修部件的备件保障度模型．通过具体算例,验证了模型的正确性;最后在备件保障度模型的基础上,给出了备件需求量模型。     

A unified approach for on-line performance evaluation of CCNs

The primary trade-off in a large computer-communication network (CCN) is the balance between computation processing capability and communication capability. Essentially, a CCN model should utilize operational optimization (rather than mathematical optimization) allowing the user (mostly a non-programme network expert) to call upon his experience and insight in dealing with real-life networks to decide upon proper direction for modifications. We propose, in this paper, a unified approach which will hopefully allow an on-line user to make real-time modifications to CCNs to obtain operationally optimal system trade-offs in such areas as network overall reliability, communication response, computation response, and optimum channel capacities etc.     

LTE网络性能智能优化平台

伴随着4G网络建设的快速发展,运营商在覆盖、干扰和容量方面均面临着更大的挑战.如何在这3个维度之间找好平衡点,最大限度提升网络性能,是中国移动各省公司在日常优化工作中面临的痛点和难点.基于这一现状,针对传统网优模式存在的问题和一线生产的重点难点,并以能输出解决方案提升网络性能为目标,我们自主研发了LTE网络性能智能优化平台.整个平台设计方案围绕LTE网络性能提升这一核心,横向涵盖了覆盖和干扰优化、以及容量和业务关联分析优化,纵向打通了网络优化全流程,实现了集中化、自动化网优的目标.本文从平台的整体架构和功能设计出发,依次介绍了各专项分析功能模块的设计和实现,最后介绍了平台在现网优化中的应用情况,以及下一步的重点研发计划.     

云计算数据中心可扩展服务器的能耗管理与优化

通过对云计算数据中心可扩展服务器功耗管理相关研究的分析,设计了云计算能耗优化管理的架构,并提出数据中心服务器节能优化机制流程图,在广泛支持的DVFS技术能耗特性基础上,提出效能优化策略,通过转移高负载,经转化高负载的主机成为低负载的主机,实现云计算数据中心可扩展服务器能耗管理的优化。采用内存为12 GB的12台主机和1台处理能力为3 200 MIPS的处理器组成数据中心进行仿真实验,结果表明,提出的效能优化策略比无迁移策略节省能耗约24.33%,比DVFS策略节省能耗约15.23%,且综合性能要优于其他两个策略,采用提出的策略可进行云计算数据中心扩展服务器能耗使用的优化。     

OASM: An overload‐aware workload scheduling method for cloud computing based on biogeographical optimization

With the daily increase in the number of cloud users and the volume of submitted workloads, load balancing (LB) over clouds followed by a reduction in users' response time is emerging as a vital issue. To successfully address the LB problem, we have optimized workload distribution among virtual machines (VMs). This approach consists of two parts: Firstly, a meta‐heuristic method based on biogeographical optimization for workload dispatching among VMs is introduced; secondly, we propose an innovative heuristic algorithm inspired by the “Banker algorithm” that runs in core scheduler to control and avoid VM overloads. The combination of these two (meta‐)heuristic algorithms constitutes an LB approach through which we have been able to reduce the value of the makespan to a reasonable time frame. Moreover, an information base repository (IBR) is introduced to maintain the online processing status of physical machines (PMs) and VMs. In our approach, data stored in IBR are retrieved when needed. This approach is compared with well‐known (non‐)evolutionary approaches, such as round‐robin, max‐min, MGGS, and TBSLB‐PSO. Experimental results reveal that our proposed approach outperforms its counterparts in a heterogeneous environment when the resources are smaller than the workloads. Moreover, the utilization of physical resources gradually increases. Therefore, optimal workload scheduling, as well as the lack of overload occurrence, results in a reduction in makespan.     

云计算环境下虚拟机在线迁移优化

在线迁移在云计算平台下已广泛使用,虚拟机内存迁移主要采用的是预拷贝算法.传统预拷贝算法在迭代过程中会将脏页反复重传导致迁移时间较长,针对这一问题文章提出了在拷贝过程中增加脏页预测算法和内存压缩算法相结合的方法.脏页预测算法是采用统计方法,依据内存页历史操作情况对改动频繁的内存页进行标记,根据标记决定是否在最后一轮内存迁移中传输,为了减少迁移传输量,先通过内存压缩算法将其压缩,然后再传输.实验结果表明,改进后的方法有效地降低了停机时间和迁移时间,提高了迁移效率,达到了更快迁移的目的.     

An availability evaluation for computer communication networks

Among the various availability criteria for computer communication networks, based on strictly topological considerations or taking into account also the traffic characteristics, we propose, as a network performance measure, the probability that an assigned fraction of the stability throughput can be carried out with average packet and message time delays not greater than specified values.At the aim, an algorithm for the availability analysis has been developed; given the topological and physical characteristics of the network, the external traffic statistics at the network node pairs and the routing procedure, the network availability and the average loss of throughput due to single failures of lines or switching nodes are evaluated with regard, as an example, to a simple five node network.     

Composite performance and availability analysis of wirelesscommunication networks

With the increasing popularity of wireless communication systems, customers are expecting the same level of service, availability, and performance from the wireless communication networks as the traditional wire-line networks. The traditional pure performance model that ignores failure and recovery but considers resource contention generally overestimates the system's ability to perform a certain job. On the other hand, pure availability analysis tends to be too conservative since performance considerations are not taken into account. To obtain realistic composite performance and availability measures, one should consider performance changes that are associated with failure recovery behavior. A review is first given over the advances in composite performance and availability analysis. Thereafter, three techniques for composite performance and availability analysis are discussed in detail through a queueing system in a wireless communication network. Numerical results show that an approximate model based on a framework originally proposed by Bobbie and Trivedi (1990) provides remarkably accurate predictions on system performance     

A high performance architecture for computing the time-frequency spectrum

To solve the problem of detecting and displaying the changes in the spectra of nonstationary signals, there are two possible approaches. Either one uses the same estimators as for the stationary signals, but one approach uses shorter-length data blocks during which the signal is assumed to be stationary, and the other one uses the same length data and applies a time-varying spectrum estimator that accounts for the nonstationarity. A time-varying spectrum estimator called a time-varying correlogram (TVC) is a well-known estimator of the time-frequency spectrum of a nonstationary signal. In this paper, a high performance VLSI architecture for computing TVC is proposed.     

An integrated testbed for wireless multimedia computing

A testbed has been constructed to evaluate node architectures that support multimedia applications and services across a wireless network. Using this testbed, a low bitrate subband video compression algorithm has been prototyped in a field programmable gate array (FPGA) and evaluated for video networking across bandwidth-limited RF channels. A radio interface has been prototyped in an FPGA and a common applications programming interface (API) has been developed to allow experimentation with multiple radios. This testbed has been used to evaluate node performance under two different wireless applications: 1) simultaneous video and data networking (VTALK) and 2) TCP/IP utilities such as FTP and telnet. Based on this evaluation, the design of a battery-operated high throughput wireless multimedia node is presented.