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

2022年10月8日

couryes-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
• (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
couryes™为您提供可以保分的包课服务

## 计算机代写|云计算代写cloud computing代考|Characteristics of Proposed Algorithm

SRTDVMC attempts to fit the largest VM in terms of VMRT from the smallest PM in terms of SPMRT into the next smallest possible PM. Such consolidation approach shortens the SPMRT of source PM without raising the SPMRT of destination $\mathrm{PM}$, resulting into decreased energy consumption. Additionally, selecting the next smallest possible PM as destination PM ensures that the PM is accomplishing largest possible jobs before moving into sleep state or turned off state. Consequently, remaining workload for the existing active PMs becomes lower, which aids in energy consumption minimization. Furthermore, lesser remaining workload for existing active PMs increases the likelihood that upcoming workload can be served by these active PMs without turning on PMs, which are in lower energy consumption state, for instance, sleep or turned off state. Hence, energy consumption minimization is complemented.

One critical aspect of SRTDVMC is that both rise of energy in potential destination PM (i.e., cost) and drop of energy in potential source PM (i.e., benefit) is checked prior any potential VM migration. VMs from $U$-UPMs are migrated only if the net energy gain (i.e., energy drop – energy rise) is positive, which limits the number of VM migrations and improves QoS without compromising energy efficiency. Hence, SRTDVMC can concurrently satisfy both objective functions (3.10) and (3.11). Furthermore, SRTDVMC smartly selects destination PMs ensuring that the increased energy consumption of potential destination $U-U P M$ does not outweigh the reduced energy consumption of potential source $U-U P M$. It aids to uphold the energy efficiency of the solution regardless of the drastic rise of state-of-the-art PMs’ energy consumption causing declined energy efficiency at utilization level beyond $70 \%$. Thus, SRTDVMC encounters the lack of energyefficiency issue in the presence of state-of-the-art PMs as experienced with existing DVMC algorithms. As a result, SRTDVMC is robust against underlying PMs’ change of energy-efficient characteristics with varying load.

## 计算机代写|云计算代写cloud computing代考|Experimental Setup

Performance of RTDVMC [26] has been evaluated through CloudSim [24]. Since, performance of $S R T D V M C$ has been compared with $R T D V M C$, therefore, we have modelled and simulated a cloud environment in CloudSim [24], which we have used to simulate SRTDVMC algorithm under different workload scenarios. For fair comparison, both algorithms have been simulated using same environment with respect to the characteristics of CDC, VM, PM and energy module. The simulated CDC consists of 800 heterogeneous PMs. Three different modern generation of PMs, such as Dell PowerEdgeR940 (Intel Xeon Platinum 8180, 112 cores $\rightarrow$ । $25,000 \mathrm{MHz}, 384 \mathrm{~GB}$ ) [30], HP ProLiant DL560 Gen10 (Intel Xeon Platinum 8180, 112 cores $\rightarrow$ 125,000 MHz, $384 \mathrm{~GB}$ ) [31], and HP ProLiant ML350 Gen10 (Intel Xeon Platinum 8180,28 cores $\rightarrow$ । $25,000 \mathrm{MHz}, 192 \mathrm{~GB}$ ) [32] have been used. Each server is provided with $1 \mathrm{~GB} / \mathrm{s}$ network bandwidth. The energy consumption characteristics of these servers with varying workload is articulated in Table 3.2.
The characteristics of different VM types match with the VMs used by $R T D V M C$ and correspond to Amazon EC2 instance types [48]. However, the difference between the simulated VMs and Amazon EC2 instance types is that the simulated VMs are single-core, which is explained by the fact that the workload data used for the simulations come from single-core VMs. Since, the single-core is used, the amount of RAM is divided by the number of cores for each VM type: HighCPU Medium Instance (2500 MIPS, $0.85$ GB); Extra Large Instance (2000 MIPS, $3.75 \mathrm{~GB}$ ): Small Instance (1000 MIPS, $1.7 \mathrm{~GB}$ ): and Micro Instance (500 MIPS. $613 \mathrm{MB})$

Lifetime of a VM $V_j$, aka VMRT of a VM $V_j$, denoted by $T_{V_j}$ can be different from one VM to another (i.e., heterogeneous). For further accurate estimation of the performance of both $S R T D V M C$ and $R T D V M C$ algorithms under real Cloud scenario, $T_{V_j}$ values are drawn from VMRT traces of a real Cloud, namely, Nectar Cloud. Nectar Cloud consists of over thousands of VMs across multiple data centers located in eight different cities of Australia [7]. For SRTDVMC algorithm, $T_{V_j}$ is converted into $S V M R T, S_{V_j}$ as per (3.1), using $0.05$ as the value of $\alpha$ and a uniformly distributed random variable ranging $[-1,+1]$ as $X$. For further clarity, maximum deviation of $T_{V_j}$ from $S_{V_j}$ is $\pm 5 \%$. At the outset, VMs are provided with the resources defined by the VM types. However, during the lifetime, VMs utilize less resources according to the workload data, widening opportunities for dynamic consolidation. The workload data also reflects traces of real Cloud workload traffic, originated as part of the CoMon project, a monitoring infrastructure for PlanetLab [49]. For both RTDVMC and SRTDVMC, upper utilization threshold, $\theta_{\max }$ is considered as $80 \%$. With every workload scenario, a DVMC algorithm has been run twice to generate mean CDC energy consumption and mean total number of VM migration by that DVMC algorithm under such workload scenario. Each time, the simulation has been run until $24 \mathrm{~h}$ CloudSim simulation clock time.

# 云计算代考

## 计算机代写|云计算代写cloud computing代考|提出的算法特征

SRTDVMC尝试将VMRT方面的最大VM从SPMRT方面的最小PM拟合到下一个最小PM。这种整合方法缩短了源PM的SPMRT，而不提高目的地$\mathrm{PM}$的SPMRT，从而降低了能源消耗。此外，选择下一个最小的PM作为目标PM可以确保PM在进入睡眠状态或关闭状态之前完成最大的工作。因此，现有活动pm的剩余工作负载变得更低，这有助于将能源消耗最小化。此外，现有活动pm剩余工作负载的减少增加了这些活动pm在不打开pm的情况下处理即将到来的工作负载的可能性，这些pm处于较低的能耗状态，例如睡眠或关闭状态。因此，能源消耗最小化得到了补充 SRTDVMC的一个关键方面是，在任何潜在VM迁移之前，都要检查潜在目的地PM的能量上升(即成本)和潜在源PM的能量下降(即收益)。仅当净能量增益(即能量下降-能量上升)为正时，才迁移$U$ – upm中的VM，这限制了VM迁移的数量，并在不影响能源效率的情况下提高了QoS。因此，SRTDVMC可以同时满足(3.10)和(3.11)两个目标函数。此外，SRTDVMC巧妙地选择目的地pm，确保潜在目的地$U-U P M$增加的能源消耗不会超过潜在源$U-U P M$减少的能源消耗。它有助于保持解决方案的能源效率，而不考虑最先进的pm的能源消耗的急剧上升，导致在利用率水平上的能源效率下降，超过$70 \%$。因此，SRTDVMC在现有DVMC算法所经历的最先进的pm存在时，会遇到缺乏能源效率的问题。结果表明，SRTDVMC对底层pm的节能特性随负载变化具有较强的鲁棒性

## 计算机代写|云计算代写cloud computing代考|Experimental Setup

. RTDVMC[26]的性能已经通过CloudSim[24]进行了评估。由于$S R T D V M C$的性能已经与$R T D V M C$进行了比较，因此，我们在CloudSim[24]中建模并模拟了一个云环境，我们使用CloudSim[24]模拟了不同工作负载场景下的SRTDVMC算法。为了便于比较，本文针对CDC、VM、PM和能源模块的特性，在相同的环境下对两种算法进行了仿真。模拟的CDC由800个异质pm组成。三种不同的现代一代pm，如戴尔PowerEdgeR940(英特尔至强白金8180,112核$\rightarrow$ ।$25,000 \mathrm{MHz}, 384 \mathrm{~GB}$) [30]， HP ProLiant DL560 Gen10 (Intel Xeon铂金8180,112核$\rightarrow$ 125000 MHz, $384 \mathrm{~GB}$)[31]，和HP ProLiant ML350 Gen10 (Intel Xeon铂金8180,28核$\rightarrow$ ।$25,000 \mathrm{MHz}, 192 \mathrm{~GB}$)[32]已被使用。每个服务器提供$1 \mathrm{~GB} / \mathrm{s}$网络带宽。表3.2列出了不同工作负载下这些服务器的能量消耗特征。

## 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。