一种基于预测的动态负载均衡模型及算法研究

提出了一种基于负载预测的动态负载均衡方法，它改变了一般动态负载均衡方法负载信息采集过时的情况，有效解决了负载迁移的抖动问题，提高了平衡系统的性能。给出了该方法的实现模型、算法，并进行了性能分析，最后给出了实验结果。     

基于退火策略的分布式资源负载均衡算法

针对分布式资源导致的访问热点等一系列问题,建立了一个用于分布式资源退火的处理模型.根据该模型,提出了基于退火策略的分布式资源负载均衡算法;该算法通过访问分类、定向扩散等方法提高了系统性能.性能及试验分析表明,该算法能够减少系统内部通信量,抑制资源扩散的抖动现象等.     

一种基于深度学习的云平台弹性伸缩算法

为了满足性能要求，降低资源消耗，研究人员提出了许多伸缩调度的算法和方案。但是，它们中的大多数只作用在服务器或应用程序的当前状态，无论是资源实际的调度效果还是算法方案的适用性上都受到了影响和限制。本文提出一种基于长短期记忆网络和BP神经网络的面向应用的弹性伸缩算法。该算法包括工作负载预测模型、响应时间预测模型和资源调整策略模型，能够对云计算应用的工作负载和响应时间进行预测并给出合适的资源调度策略。为了提高工作负载预测的准确度，本文将卷积运算和长短期网络结合起来，更好地提取数据特征并进行准确地预测。而为了提高模型收敛速度,并有效避免模型过拟合的问题，本文则在BP神经网络中使用批标准化运算。在验证实验中，该算法工作负载预测的平均绝对百分误差降低到3.4×10-4，响应时间预测和调度策略模型也达到了不错的效果。在实际平台运行中，该弹性伸缩算法还能够根据Docker容器云平台实际需要提供合适的计算资源调度策略。实验结果表明，相比较其他模型，该弹性伸缩算法在工作负载预测和云平台计算资源调整方面具有较好的性能。     

基于改进蚁群算法的网格资源调度研究

为进一步掌握网格资源动态运行状态,以便合理调度网格资源,提高任务执行效率,提出了一种基于改进蚁群算法的网格资源调度策略。该算法引入了一个网格资源空闲所需时间向量F,通过向量F动态调整网格资源负载情况,达到快速实现遥感资源空间检索的目的。从仿真实验结果可以看出,改进蚁群算法比蚁群算法和其他算法更优,网格资源的利用效率更高。     

视频网格动态负载加权平滑均衡算法

为实现网格中视频资源服务的动态负载均衡,对一种动态负载加权均衡算法进行改进。利用监测与发现系统收集每台视频服务器的CPU利用率等主要负载参数,运用上述参数加权得到综合负载,对相邻时刻的负载做平滑处理以避免调度抖动。通过比较平滑后的动态负载值与服务器综合负载阈值进行动态调度,改变相应节点的负载,避免视频服务器间的负载失衡。实验结果表明,该算法能有效降低系统平均服务延迟时间并提高吞吐量,从而提升视频资源网格服务的整体性能。     

一种多维集合划分负载均衡资源优化分配算法

针对多处理群集系统中多个任务处理需要的资源和多个处理节点能够提供的m维资源间的合理匹配问题，给出了多维集合划分问题的优化模型，定义了资源均衡度函数，提出多维集合划分负载均衡资源优化分配算法，通过该算法可以得到资源匹配NP问题的较优解。实验结果表明，该算法具有较好的实用性和可行性，比传统的启发式算法效率高。     

基于服务类型的动态反馈负载均衡算法*

针对现有静态和动态负载均衡算法往往存在计算服务节点负载过程中引用特征信息过少,或忽视不同类型服务对于节点负载的影响等问题,提出了一种基于服务类型的动态反馈算法。该算法统计各节点的多种负载信息,通过NECP协议实现动态反馈,并引入负载权重向量和负载能力向量计算节点的综合负载。算法在实际的仿真环境中得到了验证,说明了具有可用性和优越性。     

一种对等网络负载平衡算法的研究

对等网络正在成为网络应用研究的一个新的热点,负载平衡技术是其中的一个重要问题。该文提出了一种基于局部负载扩散思想的对等网络负载平衡算法,对相应的负载评估标准、负载分散原则等内容进行了详细描述,最后给出了仿真试验结果。试验数据表明,该算法具有较低的系统开销、较小的响应延迟,减少了系统抖动现象的产生。     

一种改进的动态告警负载均衡算法

负载均衡通过将大量的并发访问请求转发到多个服务器分别进行处理,以提高web集群的整体吞吐量.现有的负载均衡算法存在着引用负载因素过少、负载计算过于复杂、节点的负载抖动现象过于严重等问题.本文提出一种改进的动态告警负载均衡算法,基于请求类型、节点工作能力和实时负载值来确定转发目标.该方案实时监测各节点的负载状态并实施周期性反馈与动态告警,既保证了负载信息的实时性与有效性,又减少了负载计算量;采用RED方法校正负载状态的判定,避免了因为负载状态的误判而引发的负载不均,从而减小了负载抖动现象.模拟实验结果表明,改进算法增加了web集群系统的吞吐量,并且明显改善了负载均衡度.     

一种基于QoS的云负载均衡机制的研究

提出一种基于QoS的云负载均衡机制,即:构建QoS模型和云资源模型;建立资源度量与QoS属性之间的映射;对虚拟机实例负载状况和虚拟机集群资源利用状况进行量化评估;感知用户的QoS并对比所监控的云节点的资源度量情况,根据对比结果,通过任务调度算法和弹性伸缩算法分别实现任务的分发和虚拟机集群的弹性伸缩,最终达到优化的负载均衡的目的.通过模拟试验,结果表明本方法与Round robin算法相比,有更好的负载均衡效果.     

Failure-aware resource provisioning for hybrid Cloud infrastructure

Hybrid Cloud computing is receiving increasing attention in recent days. In order to realize the full potential of the hybrid Cloud platform, an architectural framework for efficiently coupling public and private Clouds is necessary. As resource failures due to the increasing functionality and complexity of hybrid Cloud computing are inevitable, a failure-aware resource provisioning algorithm that is capable of attending to the end-users quality of service (QoS) requirements is paramount. In this paper, we propose a scalable hybrid Cloud infrastructure as well as resource provisioning policies to assure QoS targets of the users. The proposed policies take into account the workload model and the failure correlations to redirect users’ requests to the appropriate Cloud providers. Using real failure traces and a workload model, we evaluate the proposed resource provisioning policies to demonstrate their performance, cost as well as performance–cost efficiency. Simulation results reveal that in a realistic working condition while adopting user estimates for the requests in the provisioning policies, we are able to improve the users’ QoS about 32% in terms of deadline violation rate and 57% in terms of slowdown with a limited cost on a public Cloud.     

一种云环境下的主机负载预测方法

云计算是一种基于信息网络的计算模式和服务模式,它将信息技术资源以服务方式动态、弹性地提供给用户,使用户可以按需使用。由于受到主机的启动时间、资源分配时间以及任务调度时间等因素的影响,在云环境下提供给用户的服务存在时延问题。因此,工作负载预测是云环境下一种重要的能源优化的方式。此外,由于云中工作负载的变化具有十分大的波动性,因此增加了预测模型的预测难度。提出了一种基于自回归模型和Elman神经网络的预测模型(Hybrid Auto Regressive Moving Average model and Elman neural network,HARMA-E),其使用ARMA模型进行预测,再使用ENN模型对ARMA模型的误差进行预测,通过修正ARMA的输出值得到最终的预测值。仿真实验结果表明,该预测模型能够较好地提升主机负载预测值的准确度。     

异构集群系统中一种基于资源的负载平衡算法的设计与模拟

本文指出了一种基于资源使用率和向量负载指数的,采用进程迁移机制的负载平衡算法,并通过踪迹驱动的方法进行了大量的模拟和分析。     

Optimization of virtual resource management for cloud applications to cope with traffic burst

Being the latest computing paradigm, cloud computing has proliferated as many IT giants started to deliver resources as services. Thus application providers are free from the burden of the low-level implementation and system administration. Meanwhile, the fact that we are in an era of information explosion brings certain challenges. Some websites may encounter a sharp rising workload due to some unexpected social concerns, which make these websites unavailable or even fail to provide services in time. Currently, a post-action method based on human experience and system alarm is widely used to handle this scenario in industry, which has shortcomings like reaction delay. In our paper, we want to solve this problem by deploying such websites on cloud, and use features of the cloud to tackle it. We present a framework of dynamic virtual resource management in clouds, to cope with traffic burst that applications might encounter. The framework implements a whole work-flow from prediction of the sharp rising workload to a customized resource management module which guarantees the high availability of web applications and cost-effectiveness of the cloud service providers. Our experiments show the accuracy of our workload forecasting method by comparing it with other methods. The 1998 World Cup workload dataset used in our experiment reveals the applicability of our model in the specific scenarios of traffic burst. Also, a simulation-based experiment is designed to indicate that the proposed management framework detects changes in workload intensity that occur over time and allocates multiple virtualized IT resources accordingly to achieve high availability and cost-effective targets.     

Evaluating approaches to resource demand estimation

Resource demands are a key parameter of stochastic performance models that needs to be determined when performing a quantitative performance analysis of a system. However, the direct measurement of resource demands is not feasible in most realistic systems. Therefore, statistical approaches that estimate resource demands based on coarse-grained monitoring data (e.g., CPU utilization, and response times) have been proposed in the literature. These approaches have different assumptions and characteristics that need to be considered when estimating resource demands. This paper surveys the state-of-the-art in resource demand estimation and proposes a classification scheme for estimation approaches. Furthermore, it contains an experimental evaluation comparing the impact of different factors (monitoring window size, number of workload classes, load level, collinearity, and model mismatch) on the estimation accuracy of seven different approaches. The classification scheme and the experimental comparison helps performance engineers to select an approach to resource demand estimation that fulfills the requirements of a given analysis scenario.     

Establishment of overall workload assessment technique for various tasks and workplaces

A model for assessing workloads called overall workload level (OWL) was developed by introducing linguistic variable sets and applying the analytic hierarchy process (AHP) to estimate the external workload imposed on a human operator in man–machine systems. To do this, a five-point linguistic variable set scale was constructed and their hierarchical prioritization procedures were set up. The task and workplace variables (e.g., physical, environmental, postural, and mental job demand workloads) which can obtain the operator's perception of workload are selected as workload factors and the AHP technique is used to collect different weights. Finally, OWL is calculated using a computer-assisted system to determine the level of overall workload impinged on an operator. The OWL was implemented in an actual industrial environment from a physiological and epidemiological viewpoint to determine the validity of the model. Furthermore, the results obtained by applying OWL were compared to the results obtained by applying the overall workload (OW) of the NASA task load index (TLX). The results show that there is a close linear relationship among the physiological measurements, the severity of injury and illness rates, OW, and OWL. Thus, this approach can be used for problem identification and for solving widespread occupational workloads.

Relevance to industry

The determination of workloads imposed on a human operator plays an important role in designing and evaluating an existing man–machine system. Therefore, a model for assessing workloads was developed to estimate the external workload imposed on a human operator in man–machine systems. This model can be used for problem identification and for solving widespread occupational workload.     

A hybrid wavelet decomposer and GMDH-ELM ensemble model for Network function virtualization workload forecasting in cloud computing

Nowadays Network function virtualization (NFV) has drawn immense attention from many cloud providers because of its benefits. NFV enables networks to virtualize node functions such as firewalls, load balancers, and WAN accelerators, conventionally running on dedicated hardware, and instead implements them as virtual software components on standard servers, switches, and storages. In order to provide NFV resources and meet Service Level Agreement (SLA) conditions, minimize energy consumption and utilize physical resources efficiently, resource allocation in the cloud is an essential task. Since network traffic is changing rapidly, an optimized resource allocation strategy should consider resource auto-scaling property for NFV services. In order to scale cloud resources, we should forecast the NFV workload. Existing forecasting methods are providing poor results for highly volatile and fluctuating time series such as cloud workloads. Therefore, we propose a novel hybrid wavelet time series decomposer and GMDH-ELM ensemble method named Wavelet-GMDH-ELM (WGE) for NFV workload forecasting which predicts and ensembles workload in different time-frequency scales. We evaluate the WGE model with three real cloud workload traces to verify its prediction accuracy and compare it with state of the art methods. The results show the proposed method provides better average prediction accuracy. Especially it improves Mean Absolute Percentage Error (MAPE) at least 8% compared to the rival forecasting methods such as support vector regression (SVR) and Long short term memory (LSTM).     

Microgravity simulation: physical and psychological workload evaluation tests in an underwater environment

As weightlessness is not completely reproducible on Earth, usability evaluation of space systems is often simulated through tests in an aquatic environment. A Neutral Buoyancy Facility test programme was organized in a special pool to simulate Extra-Vehicular Activities on the Columbus module of the future International Space Station with the aim of assessing various aspects of crew interface design.This study was designed to evaluate workload using visibility, accessibility and operability tests. Diving workload was determined through basic physiological measurements, such as pulmonary ventilation and heart rate during underwater operations.As anxiety can influence physiological processes, and consequently also the workload evaluation determined through these parameters, we developed an evaluation methodology to investigate the anxiety level based on a specific questionnaire submitted to all subjects before and after the dives.Heart rate increased in underwater work to a value approximately 50% larger than the value obtained in the resting condition while sitting outside the pool. This increase in heart rate was accompanied by an increase in pulmonary ventilation of 200% larger than the value recorded in the rest condition while sitting outside the water. The extent of these increases was notable in all the test subjects, who varied in age and stature.Recorded values of workload, heart rate and pulmonary ventilation were evaluated on the basis of Christensen's (Arbeitsphysiol. 14 (1950) 251) and Wells’ (J. Appl. Physiol. 10 (1957) 51) classifications. Through this analysis it was possible to determine that the workload, indicated by performance on our neutral buoyancy tests, corresponds to moderate physiological work.For test subjects, anxiety related to underwater performance was light. Among the causes of anxiety all the subjects indicated the lack of confidence with neutral buoyancy tests and a feeling of lack of safety, typical of aquatic environments.We can conclude that context did not produce considerable psychological effects, and consequently that the psychological load did not influence heart rate and pulmonary ventilation values that can therefore be directly related to task workload.     

基于粒子群算法的煤气化过程LS_SVM预测模型

针对最小二乘支持向量机参数选择对模型性能的重要影响,并且以往的参数优选方法效果差且耗时长这一问题,提出基于粒子群算法优化最小二乘支持向量机预测模型.该模型用最小二乘支持向量机理论建立,用粒子群算法优化模型参数.论文将此模型用于预测评价固定床煤气化气化效果的三个主要性能指标(气体热值、气化效率、气体产率),通过现场实际数据仿真结果表明,该算法有效地提高了模型预测精度,验证了此模型的可靠性和可用性.