## 计算机代写|云计算代写cloud computing代考|CS4740

statistics-lab™ 为您的留学生涯保驾护航 在代写云计算cloud computing方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写云计算cloud computing代写方面经验极为丰富，各种代写云计算cloud computing相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• Advanced Probability Theory 高等概率论
• Advanced Mathematical Statistics 高等数理统计学
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础

## 计算机代写|云计算代写cloud computing代考|Static Power Management

The static power management can be executed on both hardware and software levels. Leakage currents in any active circuits cause static power consumption at the hardware level [7]. SPM uses hardware components such as CPU, memory, disk storage, network devices, and power supply unit efficiently. It consists of all applied optimization methods during design time at logic, circuit, and architectural levels that will he explained in the following section.

• Logic level optimization: At this level, optimization methods attempt to optimize the power of switching activity in both sequential and combinational circuits. Minimizing the switching capacitance improves the dynamic power consumption straightly by reducing the energy per transition on each logic device $[7,41]$.
• Circuit level optimization: Significant challenges at this optimization level are based on efficient pipelining and interconnections between stages and components. Pipelining technique is regularly used to boost throughput in highperformance designs at the expense of reducing energy efficiency, contributing to increasing area and execution time [41].
• Architectural level optimization: Methods include the system’s design considering power optimization technique at an architectural level [7]. Power savings are typically accomplished at the architectural level by optimizing the system components’ balance to prevent wasting power.[41].

Besides the optimization at the hardware level, considering the SPM at the software level is also essential. Even with robust hardware design, it is crucial to be careful about software design inasmuch as weak design conduces to loss of power and performance, even with perfectly designed hardware. Thus, the code generation, the instructions used in the code, and the order of these instructions must be carefully selected, as they affect performance as well as power consumption.

## 计算机代写|云计算代写cloud computing代考|Dynamic Power Management

This section describes our taxonomy at the dynamic power management level, as shown in Fig. 2.3. DPM is categorized into three levels, including hardware, software, and hybrid. There are various kinds of optimization methods at both the hardware and software levels $[7,42]$. At the hardware level, we can imply techniques such as DVFS, DCD, and sleep states. In addition, the techniques at the software level aree classified into virtualization, migration, consolidation, plus containerization. The dynamic power consumption is induced by the high usage of hardware components (such as CPU, storage, and network devices) and the circuits’ activity. The main reason enabling dynamic power consumption pertain to both system’s components deactivation and tuning the circuit activity. Dynamic power consumption can be reached through different techniques including: (1) diminishing the switching activity, (2) decreasing the physical capacitance that relies on low-level design parameters such as transistors’ sizes, (3) ebbing the supply voltage, and (4) lessening the clock frequency [7].

DPM improves energy consumption by using knowledge gathered from current resources in the system and the workload of applications running in the system [7, 43]. DPM techniques allow dynamic adjustment of power states to occur based on current system loads. It predicts the best action in the future using the data obtained from the system and according to the system’s requirements. DPM techniques are categorized into hardware and software levels. There is another level in our taxonomy, namely hybrid, in which both hardware and software techniques are simultaneously utilized.

1. Hardware-level approaches
DPM techniques applied at the hardware level reconfigure the system dynamically by adopting methodologies to fulfill the requested services with the minimum number of active components or the minimum load on such components [43]. The DPM techniques at a hardware level can optionally turn off the idle system components or reduce the useless ones’ performance. It is also possible to exchange some components, containing CPU, between either active or idle modes to save energy. The hardware DPM techniques vary for different hardware components, yet usually, they are splitted into dynamic component deactivation (DCD) and dynamic performance scaling (DPS) [44].

## 计算机代写|云计算代写cloud computing代考|.静态电源管理

• 逻辑级优化:在这个级别上，优化方法试图优化顺序电路和组合电路中开关活动的功率。最小化开关电容通过减少每个逻辑器件上的每次跃迁能量直接提高了动态功耗$[7,41]$。
• 电路级优化:在这一优化级别上的重大挑战是基于级和组件之间的高效流水线和互连。流水线技术经常被用于提高高性能设计的吞吐量，以降低能源效率为代价，有助于增加面积和执行时间[41]。
• 体系结构级优化:方法包括体系结构级考虑功率优化技术的系统设计[7]。节能通常是通过优化系统组件的平衡在架构级实现的，以防止浪费电力除了硬件层面的优化，考虑软件层面的SPM也是必不可少的。即使有健壮的硬件设计，对软件设计也要小心谨慎，因为即使有完美设计的硬件，薄弱的设计也会导致功率和性能的损失。因此，必须仔细选择代码生成、代码中使用的指令以及这些指令的顺序，因为它们影响性能和功耗
计算机代写|云计算代写cloud computing代考|Dynamic Power Management .本节描述了我们在动态电源管理级别的分类，如图2.3所示。DPM分为三个级别，包括硬件、软件和混合。在硬件和软件层面都有各种各样的优化方法$[7,42]$。在硬件级别，我们可以使用DVFS、DCD和睡眠状态等技术。此外，软件级别的技术可以分为虚拟化、迁移、整合和容器化。动态功率消耗是由硬件组件(如CPU、存储和网络设备)和电路活动的高使用率引起的。实现动态功耗的主要原因与两个系统组件的失活和调优电路活动有关。动态功耗可以通过不同的技术实现，包括:(1)降低开关活性，(2)降低物理电容(依赖于低水平的设计参数，如晶体管的尺寸)，(3)降低电源电压，(4)降低时钟频率[7]DPM通过使用从系统中当前资源收集的知识和系统中运行的应用程序的工作负载来提高能耗[7,43]。DPM技术允许根据当前系统负载动态调整电源状态。它利用从系统中获得的数据，根据系统的需求，预测未来的最佳行动。DPM技术分为硬件级和软件级。在我们的分类法中还有另一个层次，即混合分类法，即同时使用硬件和软件技术硬件级方法应用于硬件级的DPM技术通过采用各种方法动态地重新配置系统，以实现所请求的服务，使用最少的活动组件数量或这些组件上的最小负载[43]。硬件级别的DPM技术可以选择关闭空闲的系统组件或降低无用的系统组件的性能。也可以在活动或空闲模式之间交换一些组件，包括CPU，以节省能源。硬件DPM技术因硬件组件的不同而不同，但通常分为动态组件去激活(DCD)和动态性能缩放(DPS) [44].

## 有限元方法代写

tatistics-lab作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。

## 计算机代写|云计算代写cloud computing代考|CS5412

statistics-lab™ 为您的留学生涯保驾护航 在代写云计算cloud computing方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写云计算cloud computing代写方面经验极为丰富，各种代写云计算cloud computing相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• Advanced Probability Theory 高等概率论
• Advanced Mathematical Statistics 高等数理统计学
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础

## 计算机代写|云计算代写cloud computing代考|Proposed Taxonomy for Energy-Aware Resource

In this section, we present our proposed taxonomy for energy-aware resource management solutions in cloud environments, as shown in Fig. 2.2. In our proposed taxonomy, we consider four items at the highest level. The first level pertains to the goals of energy-efficient resource management in cloud environments. As can be seen, the second level goes back to the dynamism of resource management technique. The third level is the considered type of workload, including arbitrary, High-Performance Computing (HPC), batch, and real-time applications. Finally, the fourth level is the type of resources that are classified into active and passive. The details of this taxonomy are described in the following subsection.

As energy-efficient resource management approaches play a crucial role in cloud data centers these days, there are various kinds of goals in this way. To clarify, Fig. $2.2$ summarizes the components of goals. Indeed, we have considered five targets for this component, such as minimizing power consumption, maximizing performance, load balancing, meeting power budget, plus maximizing business profit. We have also regarded four performance metrics, including response time, SLA violation, throughput, and delay. First and foremost, data centers have significant power cost; thereby, it is an essential requirement for data centers’ operation to meet the power budget coming from the limitation for power usage and observing this limitation [25].

Due to load imbalance, some of the data center resources may become overloaded or underloaded, which leads to performance degradation and resource wastage. Load balancing conduces to maximize resource utilization and achieve the desired QoS in the cloud by employing optimal resource allocation and workload distribution approaches at both schedule and runtime.

• Batch processing: Theoretically, batch processing is a processing mode when a sequence of jobs are executed on a batch of inputs [27]. Analyzing data on a large scale and batch processing occurs by utilizing data centers and some distributed and computing frameworks such as Map-Reduce and Hadoop [4]. Map-Reduce programming paradigm is the most practical and efficient solution for batch processing of big data [28].
• HPC: In the early 1990s, clusters of computers became famous in HPC environments owing to their low cost compared to custom supercomputers and mainframes [29]. Also, HPC computers generally take advantage of open source operating systems such as Linux. In the early 2000s, grid computing was linked to the HPC community as a consequence of need to run parallel programs even larger than that was normal in grid environments. Grids provide powerful resources operated by independent administrative domains to users [30]. In the late 2000 s, cloud computing was quickly growing its adolescent level and reputation, and studies started to appear on the viability of executing HPC applications on remote cloud resources [31, 32]. HPC applications are resource-intensive scientific workflows (in terms of data, computation, and communication) that have usually aimed at Grids and customary HPC platforms like super-computing clusters [33]. Both the size and number of HPC data centers have overgrown in recent years, which conduces to an exponential increase in power drastically [34].
• Real-time application: With improved cloud computing infrastructure, realtime computing can be accomplished on cloud infrastructure [35]. In most of the real-time cloud applications, the processing is executed on remote cloud computing nodes. As we meet many real-time systems around us, Cloud’s support plays a crucial role in the real-time system [36]. Their application ranges from small mobile phones to huge industrial controls and from a mini pacemaker to larger nuclear plants. A usual real-time, such as financial analysis, distributed databases, or image processing, includes multiple real-time applications or subtasks service [37]. Real-time systems are implemented by several simultaneous tasks requesting to access hardware resources [24].

## 计算机代写|云计算代写cloud computing代考|能源感知资源的建议分类法

.

.

• HPC:在20世纪90年代早期，集群计算机因其与定制超级计算机和大型机[29]相比的低成本而在HPC环境中出名。此外，高性能计算计算机通常利用开源操作系统，如Linux。在21世纪初，网格计算与HPC社区联系在一起，因为需要运行比网格环境中正常情况下更大的并行程序。网格为用户[30]提供了由独立管理域操作的强大资源。在2000年代后期，云计算的水平和声誉迅速提高，在远程云资源上执行高性能计算应用的可行性研究开始出现[31,32]。HPC应用程序是资源密集型的科学工作流(就数据、计算和通信而言)，通常针对网格和传统的HPC平台，如超级计算集群[33]。近年来，HPC数据中心的规模和数量都急剧增长，这导致了功率的指数级急剧增长

## 有限元方法代写

tatistics-lab作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。

## 计算机代写|云计算代写cloud computing代考|ECE4150

statistics-lab™ 为您的留学生涯保驾护航 在代写云计算cloud computing方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写云计算cloud computing代写方面经验极为丰富，各种代写云计算cloud computing相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• Advanced Probability Theory 高等概率论
• Advanced Mathematical Statistics 高等数理统计学
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础

## 计算机代写|云计算代写cloud computing代考|Motivation and Contribution

With the rapid ever-increasing demand to access diverse information and communication technology services based on the cloud service delivery model, the number of huge energy-hungry cloud data centers is increasing rapidly. Thus, these days energy consumption in cloud environments is a crucial issue. A data center will consume the energy of about 1000TWH in next ten years (2013-2025) [6]. The percentage of energy consumption by the data centers and cooling systems will reach $5 \%$ of the total energy consumption in the world. Energy consumption leads to operational cost and environmental implications such as global warming [7].

This key challenge leads to a rethink about the techniques and research strategies to lessen energy consumption as a crucial matter in the cloud environment. To overwhelm this challenge, there are various solutions that researchers have introduced; among them, resource management techniques play a significant role. Nonetheless, energy efficiency is still a challenge for future researchers [8]. Virtualization technique enables cloud providers to create multiple Virtual Machine(VM) instances on a single physical server, or multiple containers on a VM or a physical server which makes it possible to have servers with higher utilization. The dynamic consolidation of both VMs and containers through live migration is an efficient approaches for saving energy in cloud data centers [9]. Although the recent technological developments and paradigms including High-Performance Computing (HPC), containerization, exascale computing, and processing at network edge appear to yield new opportunities for cloud computing, they are also creating new challenges and demands for new approaches and research strategies.

Container technology has emerged thanks to Docker [10] which has boosted in both academia and industry. It provides a way to package an application that can be run with its dependencies and libraries isolated from other applications. Containers arose as a lightweight alternative of VMs that present better microservice architecture supports. The technology of container is strongly supported by PaaS, IaaS, and Internet Service Providers. Traditional hypervisor-based solutions are virtualized at the hardware level, while containerization provides virtualization at the operating system level. The containers interact with each other via system standard calls and they do not have any information about themselves [11]. Although VM technology needs to have an individual operating system for each VM, only one operating system can serve all containers in container technology. So, container technology provides more lightweight virtual systems which makes it possible to utilize system resources such as CPU, RAM, and network bandwidth more efficiently [10]. This happens owing to Linux kernel’s cgroups and namespaces which are used by docker. Besides, utilizing container technology considerably decreases startup time and the expected resources for each image in comparison with VM technology. To exemplify, a container requires 50 milliseconds to start, while a VM is activated in $30-40$ seconds [12].

Many Internet companies have embraced this technology and containers have become the de-facto standard for creating, publishing, and running applications. On the other side, there are still impediments and challenges in container-based virtualization demanding to be addressed, including security issues, in particular, during migration, dynamic resource allocation, and energy consumption [13].

## 计算机代写|云计算代写cloud computing代考|Related Work

Energy-efficient resource management approaches in cloud environments are a hot topic which vastly addressed by researchers. Since cloud computing’s research has advanced continuously, there is a need for a systematic review to evaluate, update, and join the existing literature. This section summarizes some of the previous works in the literature similar to our work, the result of which is shown in Table 2.1.
The authors in [15] have conducted a survey on energy-aware resource allocation in cloud data centers. They have reviewed some keywords such as virtualization, allocation of VM, energy efficiency, power consumption, as well as cloud computing. They have discussed various kinds of energy-aware system architectures for the cloud and compared energy efficiency in both traditional and virtual data centers. This chapter has further proposed a taxonomy for energy-saving methods in cloud data centers, which were studied in three levels, such as power management, resource management, and thermal management. The researchers have reviewed previous works based on VM allocation algorithms, VM selection algorithms, and Dynamic Voltage Frequency Scaling(DVFS), which conduced to energy saving. Plus, they have shown that the energy-saving approach became possible using renewable energy that plenty of recent research introduced this strategy.

In [16] the authors have presented a brief survey describing primary energyconserving techniques in the cloud environment. To add, they have classified energy consumption approaches into five categories, including energy-efficient hardware, energy-aware scheduling, consolidation, energy conservation in a cluster of servers, as well as power-efficient networks. Finally, they have evaluated a few papers based on this classification. The researchers further have focused on consolidation techniques in three levels, containing task consolidation, server consolidation, and energy-aware task consolidation.

Researchers in [24] have performed a comprehensive survey on energy-efficient computing, clusters, grids, and clouds. They have reported a number of approaches in the literature which contributed to improve energy efficiency. This chapter has proposed three taxonomies, covering such levels as scheduling, energy-efficient computing, as well as energy-efficient technique at different levels to make data center greener. Plus, [24] studied the energy efficiency of a single system and largescale cloud data centers, storage systems, and networking.

## 计算机代写|云计算代写cloud computing代考|动机和贡献

. cloud computing

## 计算机代写|云计算代写cloud computing代考|相关工作

. bat

[24]的研究人员对节能计算、集群、网格和云进行了全面的调查。他们在文献中报告了一些有助于提高能源效率的方法。本章提出了三个分类，包括调度、节能计算和不同级别的节能技术，使数据中心更绿色。此外，[24]研究了单个系统和大型云数据中心、存储系统和网络的能源效率

## 有限元方法代写

tatistics-lab作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。