Large-scale Cloud Data Centers Research Articles

A Systematic Parameter Analysis of Cloud Simulation Tools in Cloud Computing Environments

To provide various applications in various domains, a large-scale cloud data center is required. Cloud computing enables access to nearly infinite computing resources on demand. As cloud computing grows in popularity, researchers in this field must conduct real-world experiments. Configuring and running these tests in an actual cloud environment is costly. Modeling and simulation methods, on the other hand, are acceptable solutions for emulating environments in cloud computing. This research paper reviewed several simulation tools specifically for cloud computing in the literature and presented the most effective simulation methods in this research domain, as well as an analysis of a variety of cloud simulation tools. Cloud computing tools such as CloudSim, CloudSim Plus, CloudAnalyst, iFogSim, and CloudReports were evaluated. Furthermore, a parametric evaluation of cloud simulation tools is presented based on the identified parameters. Several 5-parameter tests were performed to demonstrate the capabilities of the cloud simulator. These results show the value of our proposed simulation system. CloudSim, CloudSim Plus, CloudAnalyst, iFogSim, and CloudReports are used to evaluate host processing elements, virtual machine processing elements, cloudlet processing elements, userbase average, minimum, and maximum, and cloudlet ID Start Time, Finish Time, Average Start, and Average Finish for each simulator. The outcomes compare these five simulator metrics. After reading this paper, the reader will be able to compare popular simulators in terms of supported models, architecture, and high-level features. We performed a comparative analysis of several cloud simulators based on various parameters. The goal is to provide insights for each analysis given their features, functionalities, and guidelines on the way to researchers’ preferred tools.

Task Failure Prediction in Cloud Data Centers Using Deep Learning

A large-scale cloud data center needs to provide high service reliability and availability with low failure occurrence probability. However, current large-scale cloud data centers still face high failure rates due to many reasons such as hardware and software failures, which often result in task and job failures. Such failures can severely reduce the reliability of cloud services and also occupy huge amount of resources to recover the service from failures. Therefore, it is important to predict task or job failures before occurrence with high accuracy to avoid unexpected wastage. Many machine learning and deep learning based methods have been proposed for the task or job failure prediction by analyzing past system message logs and identifying the relationship between the data and the failures. In order to further improve the failure prediction accuracy of the previous machine learning and deep learning based methods, in this article, we propose a failure prediction algorithm based on multi-layer Bidirectional Long Short Term Memory (Bi-LSTM) to identify task and job failures in the cloud. The goal of Bi-LSTM failure prediction algorithm is to predict whether the tasks and jobs are failed or completed. The trace-driven experiments show that our algorithm outperforms other state-of-art prediction methods with 93 percent accuracy and 87 percent for task failure and job failures respectively.

Diagnosing Evolution of Cloud Cluster via Spatio-temporal Trace Analysis

To analyze inner-enterprise cloud cluster performance, the role of workload analysis is of paramount interest to system designers. However, the ever-evolving nature of inner-enterprise cloud platforms such as diversity and spatio-temporal nature of workloads makes evolution diagnosing a challenging task. In this paper, we propose MuCoTrAna-Inner, an evolution diagnosing approach for a large-scale cloud data center based on comparative spatio-temporal trace analysis. Moreover, we present a case study on two representative big traces: Alibaba 2017 trace, and Alibaba 2018 trace. Novel quantitative findings along with the performance bottleneck inferences and recommendations based on workload analysis are provided. Our multifaceted analyses of the traces and new findings not only reveal interesting insights that are of interest to system designers and administrators, but also establish a new view to diagnosing the evolution of inner-enterprise cloud cluster based on trace analysis.

Enhancing the Security and Performance of Cloud for E-Governance Infrastructure

E-Governance is getting momentous in India. Over the years, e-Governance has played a major part in every sphere of the economy. In this paper, we have proposed E-MODI (E-governance model for open distributed infrastructure) a centralized e-Governance system for government of India, the implementation of this system is technically based on open distributed infrastructure which comprises of various government bodies in one single centralized unit. Our proposed model identifies three different patterns of cloud computing which are DGC, SGC and CGC. In addition, readiness assessment of the services needs to migrate into cloud. In this paper, we propose energy efficient VM allocation algorithm to achieve higher energy efficiency in large scale cloud data centers when system on optimum mode. Our objectives have been explained in details and experiments were designed to demonstrate the robustness of the multi-layered security which is an integration of High secure lightweight block cipher CSL along with Ultra powerful BLAKE3 hashing function in order to maintain information security triad.

EAMA: Efficient Adaptive Migration Algorithm for Cloud Data Centers (CDCs)

The rapid demand for Cloud services resulted in the establishment of large-scale Cloud Data Centers (CDCs), which ultimately consume a large amount of energy. An enormous amount of energy consumption eventually leads to high operating costs and carbon emissions. To reduce energy consumption with efficient resource utilization, various dynamic Virtual Machine (VM) consolidation approaches (i.e., Predictive Anti-Correlated Placement Algorithm (PACPA), Resource-Utilization-Aware Energy Efficient (RUAEE), Memory-bound Pre-copy Live Migration (MPLM), m Mixed migration strategy, Memory/disk operation aware Live VM Migration (MLLM), etc.) have been considered. Most of these techniques do aggressive VM consolidation that eventually results in performance degradation of CDCs in terms of resource utilization and energy consumption. In this paper, an Efficient Adaptive Migration Algorithm (EAMA) is proposed for effective migration and placement of VMs on the Physical Machines (PMs) dynamically. The proposed approach has two distinct features: first, selection of PM locations with optimum access delay where the VMs are required to be migrated, and second, reduces the number of VM migrations. Extensive simulation experiments have been conducted using the CloudSim toolkit. The results of the proposed approach are compared with the PACPA and RUAEE algorithms in terms of Service-Level Agreement (SLA) violation, resource utilization, number of hosts shut down, and energy consumption. Results show that proposed EAMA approach significantly reduces the number of migrations by 16% and 24%, SLA violation by 20% and 34%, and increases the resource utilization by 8% to 17% with increased number of hosts shut down from 10% to 13% as compared to the PACPA and RUAEE, respectively. Moreover, a 13% improvement in energy consumption has also been observed.

Energy, network, and application-aware virtual machine placement model in SDN-enabled large scale cloud data centers

Cloud computing has been considered a core model of elastic on-demand resource allocation using a pay-as-you go model. One of the big challenges of this environment is to provide high quality service (QoS) through efficient and stringent management of cloud data center resources. With the increasing demand for cloud based services, the traffic volume inside cloud data centers (DC) has been increased exponentially. Accordingly, and to provide high QoS, a proper scheduling mechanism has to be followed by the cloud service provider. Furthermore, accurate scheduling is necessary for advancing the problem of energy consumption and resource utilization. In this paper, we propose an optimal resource allocation and consolidation virtual machine (VM) placement model for multi-tier applications in modern large cloud DCs. The proposed model targets to optimize the DCs’ energy and communication cost that influence the overall cloud performance through Software Defined Networking (SDN) control features. To solve the formulated multi-objective optimization problem, a novel adaptive genetic algorithm is proposed. The experimental results validate the efficacy of the proposed model through extensive simulations using synthetic and real workload traces. These results show that the proposed model jointly optimizes cloud QoS as well as energy consumption.

The Power Control Method of Data Center Based on Cloud- Edge Collaboration

Data centers have the characteristics of huge energy consumption and strong power fluctuations. To solve the problem of unstable supply and demand such as frequent power fluctuations in the micro-grid of the data center using renewable energies, this paper establishes a controllable load model of server clusters based on the task migration mechanism by analyzing the various tasks and characteristics of large cloud data centers, as well as edge computing nodes. Through the collaborative optimization control of edge and cloud computing tasks, and dynamically adjusting and migrating the server cluster load, a new data center tie-line power control method is proposed. The results of calculation examples show that the method can achieve power smoothing of large-scale cloud data center tie lines through cloud-edge collaborative control.

Clustered Multicast Source Routing for Large-Scale Cloud Data Centers

The multi-tenancy concept in cloud data center (DC) networks paves the way towards advancements and innovation in the underlying infrastructure such as network virtualization. Multicast routing is essential in leveraging multi-tenancy to its full potential. However, traditional IP multicast routing is not suitable for DC networks due to the need to support a massive amount of multicast groups and hosts. State-of-the-art DC multicast routing approaches aim to overcome these scalability issues by, for instance, taking advantage of the symmetry of DC topologies and the programmability of DC switches to compactly encode multicast group information inside packets, thereby reducing the overhead resulting from the need to store the states of flows at the network switches. Although these approaches scale well with the number of multicast groups, they do not perform well with group sizes and, as a result, yield substantial traffic control overhead and network congestion. In this article, we present Bert, a scalable source-initiated DC multicast routing approach that scales well with both the number and size of multicast groups through the clustering of multicast group members where each cluster employs its own forwarding rules. Compared to the state-of-the-art approach, Bert yields much less traffic control overhead by significantly reducing packet header sizes and eliminating switch memory usage across the switches.

Towards Service Composition Aware Virtual Machine Migration Approach in the Cloud

There is a growing trend for service providers to migrate their services from local clusters to the cloud data center. When there is no single service can satisfy the functionality requirement of the end user, existing services are combined together to fulfill the requirements. The data communication between component service hosting servers imposes a heavy burden on the data center network. In this article, we seek to reduce the data center network resource consumption by designing a novel service composition aware virtual machine migration approach. First, we formulate the problem as a multi-object integer non-linear(INLP) programming problem. The problem, which can be reduced into a well-known multi-object quadratic assignment problem, is proved to be NP-hard. Second, we simplify the multiple-objects INLP formulation into an equivalent, but much simplified single object ILP formulation. Then, we prove that the simplified formulation can also lead to the optimal solutions. Finally, optimization problem solvers, such as LPSolver, are employed to solve the problem. Experimental results in a large scale cloud data center demonstrate that our method significantly reduce the network resource consumption than other approaches.

GRVMP: A Greedy Randomized Algorithm for Virtual Machine Placement in Cloud Data Centers

Cloud computing efficiency greatly depends on the efficiency of the virtual machines (VMs) placement strategy used. However, VM placement has remained one of the major challenging issues in cloud computing mainly because of the heterogeneity in both virtual and physical machines (PMs), the multidimensionality of the resources, and the increasing scale of the cloud data centers (CDCs). An inefficiency in VM placement strategy has a significant influence on the quality of service provided, the amount of energy consumed, and the running costs of the CDCs. To address these issues, in this article, we propose a greedy randomized VM placement (GRVMP) algorithm in a large-scale CDC with heterogeneous and multidimensional resources. GRVMP inspires the “power of two choices” model and places VMs on the more power-efficient PMs to jointly optimize CDC energy usage and resource utilization. The performance of GRVMP is evaluated using synthetic and real-world production scenarios (Amazon EC2) with several performance matrices. The results of the experiment confirm that GRVMP jointly optimizes power usage and the overall wastage of resource utilization. The results also show that GRVMP significantly outperforms the baseline schemes in terms of the performance metrics used.

An energy-efficient algorithm for virtual machine placement optimization in cloud data centers

Cloud providers offer computing services based on user demands using the Infrastructure as a Service (IaaS) service model. In a cloud data center, it is possible that multiple Virtual Machines (VMs) run on a Physical Machine (PM) using virtualization technology. Virtual Machine Placement (VMP) problem is the mapping of virtual machines across multiple physical ones. This process plays a vital role in defining energy consumption and resource usage efficiency in the cloud data center infrastructure. However, providing an efficient solution is not trivial due to difficulties such as machine heterogeneity, multi-dimensional resources, and large scale cloud data centers. In this paper, we propose an efficient heuristic algorithm that focuses on power consumption and resource wastage optimization to solve the aforementioned problem. The proposed algorithm, called MinPR, minimizes the total power consumption by reducing the number of active physical machines and prioritizing the power-efficient ones. Also, it reduces resource wastage by maximizing and balancing resource utilization among physical machines. To achieve these goals, we propose a new Resource Usage Factor model that manages virtual machine placement on physical machines using reward and penalty mechanisms. Simulations based on cloud user-customized VMs and Amazon EC2 Instances workloads illustrate that the proposed algorithm outperforms existing approaches. In particular, the proposed algorithm reduces total energy consumption by up to 15% for cloud user-customized VMs and by up to 10% for Amazon EC2 Instances.

Energy-driven cloud simulation: existing surveys, simulation supports, impacts and challenges

A large scale cloud data center is needed to provision various applications in different domains. As a result, power consumption is expected to increase due to huge operations and expansion of cloud data centers. Furthermore, it also intensifies environment concern. Various approaches and solutions for energy-driven cloud data center have been proposed to overcome this challenge. Testing and evaluating these solutions in large scale is costly and time consuming. Hence, simulation techniques become the preferred approach to tackle this concern. There are a few cloud simulators have been developed with different features and capabilities which can be chosen for this reason. A survey work can serve as a guideline. A few cloud simulation surveys have been done but limited survey is found for energy-driven cloud simulation. This review complements the existing surveys by considering different aspects of energy-driven cloud simulators. Therefore, this paper presents a review of existing cloud simulation surveys with several classifications. Furthermore, it provides some insights of the selected cloud simulators by emphasizing on the energy-driven simulation supports and the impact of the cloud simulators in succeeding works. This paper also highlights open and future challenges.

Embedding individualized machine learning prediction models for energy efficient VM consolidation within Cloud data centers

The fast growth in demand for utility-based IT services has lead to the formation of large scale Cloud data centers. The electrical energy consumption of these data centers results in high operational costs and carbon dioxide emissions. Cloud data centers benefit from the use of virtualization technology to reduce their energy consumption. This technology enables a Cloud data center to allocate its physical resources (CPU, memory, hard disk, network bandwidth) on demand and balance loads between their physical hosts by live migration of Virtual Machines (VMs). However, the migration of VMs can result in Service Level Agreement Violations (SLAVs) and consequently low Quality of Service (QoS). Hence, in this paper, we propose an energy aware VM consolidation algorithm that minimizes SLAVs.Dynamic VM consolidation has three stages: a) Detecting over- and under-utilized hosts; b) Selecting one or more VMs for migration from those hosts; c) Finding destination hosts for the selected VMs. Therefore, the proposed VM consolidation algorithm contains different models for each stage. For the first stage, we developed different fine-tuned Machine Learning (ML) prediction models for individual VMs to predict the best time to trigger migrations from hosts. For the second stage, we lexicographically consider migration time and host CPU usage when selecting VMs to migrate. Finally, a new method based on the Best Fit Decreasing (BFD) algorithm was developed to select a destination host for the VMs being migrated. Our algorithm was compared to a baseline VM consolidation algorithm that used Local Regression for detecting over-utilized hosts, minimum migration time for the VM selection stage and power-aware best fit for the host selection stage. The comparison demonstrated that our VM consolidation algorithm improved energy consumption and SLAVs by 26% and 50%, respectively.

Resource Prediction for Big Data Processing in a Cloud Data Center : A Machine Learning Approach

The high demand for big data applications, such as the Internet of Things (IoT), healthcare, business, and academia, as well as government, fosters the creation of large-scale cloud data centers. Cloud data centers contain thousands of physical machines (PMs), so resource management is necessary for allocating the tremendous amount of data to them. Knowing the workload demand in advance enables control of those resources, saving energy, reducing CPU and memory usage, and improving service. Workload prediction can be used to determine how many resources need to be allocated in the future. In this paper, we propose machine learning–based techniques to predict the daily operational workload. The proposed approach can predict the amount of power consumption (PC) and the number of PMs required to fulfill the demands of the cloud data center. Workload prediction accuracy varies based on the prediction methods used and the type of workload. In this work, we investigate three different methods: polynomial regression, support vector regression, and random forest regression (RFR). Considering both accuracy and computation time, results show that RFR provides the best performance, in our case, with a minimum root-mean-square error of 11.68 for PMs and 4869.08 for PC prediction. The computation time solidifies our selection with 2 seconds training time in all instances.

Asymptotically Optimal Load Balancing Topologies

We consider a system of N ~servers inter-connected by some underlying graph topology~G N . Tasks with unit-mean exponential processing times arrive at the various servers as independent Poisson processes of rate lambda. Each incoming task is irrevocably assigned to whichever server has the smallest number of tasks among the one where it appears and its neighbors in G N . The above model arises in the context of load balancing in large-scale cloud networks and data centers, and has been extensively investigated in the case G N is a clique. Since the servers are exchangeable in that case, mean-field limits apply, and in particular it has been proved that for any lambda < 1, the fraction of servers with two or more tasks vanishes in the limit as N -> ınfty. For an arbitrary graph G N , mean-field techniques break down, complicating the analysis, and the queue length process tends to be worse than for a clique. Accordingly, a graph G N is said to be N -optimal or ∞N-optimal when the queue length process on G N is equivalent to that on a clique on an N -scale or ∞N-scale, respectively. We prove that if G N is an Erdos-Rényi random graph with average degree d(N), then with high probability it is N -optimal and ∞N-optimal if d(N) -> ınfty$ and d(N) / (∞N łog(N)) -> ınfty as N -> ınfty, respectively. This demonstrates that optimality can be maintained at N -scale and ∞N-scale while reducing the number of connections by nearly a factor N and ∞N/ łog(N) compared to a clique, provided the topology is suitably random. It is further shown that if G N contains Θ(N) bounded-degree nodes, then it cannot be N -optimal. In addition, we establish that an arbitrary graph G N is N -optimal when its minimum degree is N - o(N), and may not be N -optimal even when its minimum degree is c N + o(N) for any 0 < c < 1/2. Simulation experiments are conducted for various scenarios to corroborate the asymptotic results.

Type-aware virtual machine management for energy efficient cloud data centers

To meet the growing demands on cloud services and applications, a sizeable number of large scale cloud data centers, hosting thousands of heterogeneous servers, is established by cloud service providers. The ever growth in establishing cloud data centers is accompanied by consuming enormous amounts of energy. Thus, proposing efficient management approaches to reduce the energy consumption in cloud data centers becomes a top priority for ensuring the scalability of the cloud computing architecture. In general, the main source of energy overconsumption in today’s data centers is due to the inefficient use of the physical servers’ resources, which results in poor server utilization patterns. So, the key aspect is to utilize the physical resources optimally while serving the cloud user demands. This paper investigates the design and implementation of virtual machine management strategies for energy efficient cloud data centers. Particularly, it considers the processes of virtual machine placement and virtual machine consolidation in enhancing the energy efficiency in cloud infrastructures. While addressing the virtual machine placement problem is important, virtual machine consolidation is even more important to enable continuous reorganization of the already-placed virtual machines on the least number of physical machines. This results in reducing the number of active physical machines by leveraging live virtual machine migration enabled by the virtualization concept. Moreover, since the virtual machine migration operations consume additional energy, the frequency of VM migrations needs to be limited and controlled as well. The paper presents a distributed approach to an energy-efficient dynamic virtual machine consolidation mechanism. This approach determines, based on novel algorithms, which virtual machines to migrate, and when. Then, the placement of the virtual machines selected for migration is achieved based on a generalization of the Knapsack Problem known as the Multiple Choice Knapsack Problem. The placement process suits both static and dynamic virtual machine placement. The results of the performance evaluation demonstrate that the proposed new algorithms are able to enhance the energy efficiency in cloud data centers.

5G-Based Transmission Power Control Mechanism in Fog Computing for Internet of Things Devices

Fog computing has become the revolutionary paradigm and one of the intelligent services of the 5th Generation (5G) emerging network, while Internet of Things (IoT) lies under its main umbrella. Enhancing and optimizing the quality of service (QoS) in Fog computing networks is one of the critical challenges of the present. In the meantime, strong links between the Fog, IoT devices and the supporting back-end servers is done through large scale cloud data centers and with the linear exponential trend of IoT devices and voluminous generated data. Fog computing is one of the vital and potential solutions for IoT in close connection with things and end users with less latency but due to high computational complexity, less storage capacity and more power drain in the cloud it is inappropriate choice. So, to remedy this issue, we propose transmission power control (TPC) based QoS optimization algorithm named (QoS-TPC) in the Fog computing. Besides, we propose the Fog-IoT-TPC-QoS architecture and establish the connection between TPC and Fog computing by considering static and dynamic conditions of wireless channel. Experimental results examine that proposed QoS-TPC optimizes the QoS in terms of maximum throughput, less delay, less jitter and minimum energy drain as compared to the conventional that is, ATPC, SKims and constant TPC methods.

Asymptotically Optimal Load Balancing Topologies

We consider a system of N servers inter-connected by some underlying graph topology G N . Tasks with unit-mean exponential processing times arrive at the various servers as independent Poisson processes of rate λ. Each incoming task is irrevocably assigned to whichever server has the smallest number of tasks among the one where it appears and its neighbors in G N . The above model arises in the context of load balancing in large-scale cloud networks and data centers, and has been extensively investigated in the case G N is a clique. Since the servers are exchangeable in that case, mean-field limits apply, and in particular it has been proved that for any λ < 1, the fraction of servers with two or more tasks vanishes in the limit as N → ∞. For an arbitrary graph G N , mean-field techniques break down, complicating the analysis, and the queue length process tends to be worse than for a clique. Accordingly, a graph G N is said to be N -optimal or √ N -optimal when the queue length process on G N is equivalent to that on a clique on an N -scale or √ N -scale, respectively. We prove that if G N is an Erdöo s-Rényi random graph with average degree d ( N ), then with high probability it is N -optimal and √ N -optimal if d ( N ) → ∞ and d ( N )/√ N log( N )) → ∞ as N → ∞, respectively. This demonstrates that optimality can be maintained at N -scale and √ N -scale while reducing the number of connections by nearly a factor N and √ N /log( N ) compared to a clique, provided the topology is suitably random. It is further shown that if G N contains Θ( N ) bounded-degree nodes, then it cannot be N -optimal. In addition, we establish that an arbitrary graph G N is N -optimal when its minimum degree is N - o ( N ), and may not be N -optimal even when its minimum degree is cN + o ( N ) for any 0 < c < 1/2. Simulation experiments are conducted for various scenarios to corroborate the asymptotic results.

Adaptive Energy-Aware Algorithms for Minimizing Energy Consumption and SLA Violation in Cloud Computing

In cloud computing, high energy consumption and service-level agreements (SLAs) violation are the challenging issues considering that the demand for computational power is growing rapidly, thereby requiring large-scale cloud data centers. Although, there are many existing energy-aware approaches focusing on minimizing energy consumption while ignoring the SLA violation at the time of a virtual machine (VM) selection from overloaded hosts. Also, they do not consider that the current network traffic causes performance degradation and thus may not really reduce SLA violation under a variety of workloads. In this context, this paper proposes three adaptive models, namely, gradient descent-based regression (Gdr), maximize correlation percentage (MCP), and bandwidth-aware selection policy (Bw), that can significantly minimize energy consumption and SLA violation. Energy-aware methods for overloaded host detection and VM selection from an overloaded host are necessary to improve the energy efficiency and SLA violation of a cloud data center after migrating all VM from underloaded host turn to idle host, which switch to energy-saving mode is also beneficial. Gdr and MCP are adaptive energy-aware algorithms based on the robust regression model, for overloaded host detection. A Bw dynamic VM selection policy selects VM according to the network traffic from the overloaded host under SLAs. Experimental results on the real workload traces show that the proposed algorithms reduce energy consumption while maintaining the required performance levels in a cloud data center using a CloudSim simulator to validate the proposed algorithms.

A critical survey of live virtual machine migration techniques

Virtualization techniques effectively handle the growing demand for computing, storage, and communication resources in large-scale Cloud Data Centers (CDC). It helps to achieve different resource m...