EC2 Instances Research Articles

Craftify: An Online Marketplace for Handcrafted Treasures

Abstract : THIS is the project about the cloud services provider to any of site and application and cloud storage and databases services to client Shirley Rodriguez, a researcher at the nonprofit organization, developed the website. She thought it would be valuable to share the data that she had gathered with other researchers. Shirley stores the data in a MySQL database, and the data is available through a PHP website that she built. She initially published the site through a commercial hosting company that provides limited support for technical issues security.Over the past year, Shirley’s website has grown in popularity. As a result of increased traffic, she started receiving complaints that the site is not as responsive as it used to be. She also experienced an attempted ransomware security breach. The security breach was unsuccessful, but her supervisor, Mateo Jackson, suggested that Shirley investigate new ways to host the website.Shirley heard about AWS and initially moved her website and database to an EC2 instance that runs in a public subnet. She also runs an instance of MySQL on the same EC2 instance.Shirley approached your team to make sure that her current design follows architectural best practices. She wants to make sure that she has a robust and secure website. One of your colleagues started the process of migrating the site to a more secure implementation, but they were reassigned to another project. Your tasks are to complete the implementation, make sure that the website is secure, and confirm that the website returns data from the query page. IndexTerms – E-commerce, Online Marketplace, Handcrafted Products, Artisan Empowerment, Cultural Preservation.

GenArchBench: A genomics benchmark suite for arm HPC processors

Arm usage has substantially grown in the High-Performance Computing (HPC) community. Japanese supercomputer Fugaku, powered by Arm-based A64FX processors, held the top position on the Top500 list between June 2020 and June 2022, currently sitting in the fourth position. The recently released 7th generation of Amazon EC2 instances for compute-intensive workloads (C7 g) is also powered by Arm Graviton3 processors. Projects like European Mont-Blanc and U.S. DOE/NNSA Astra are further examples of Arm irruption in HPC. In parallel, over the last decade, the rapid improvement of genomic sequencing technologies and the exponential growth of sequencing data has placed a significant bottleneck on the computational side. While most genomics applications have been thoroughly tested and optimized for x86 systems, just a few are prepared to perform efficiently on Arm machines. Moreover, these applications do not exploit the newly introduced Scalable Vector Extensions (SVE).This paper presents GenArchBench, the first genome analysis benchmark suite targeting Arm architectures. We have selected computationally demanding kernels from the most widely used tools in genome data analysis and ported them to Arm-based A64FX and Graviton3 processors. Overall, the GenArch benchmark suite comprises 13 multi-core kernels from critical stages of widely-used genome analysis pipelines, including base-calling, read mapping, variant calling, and genome assembly. Our benchmark suite includes different input data sets per kernel (small and large), each with a corresponding regression test to verify the correctness of each execution automatically. Moreover, the porting features the usage of the novel Arm SVE instructions, algorithmic and code optimizations, and the exploitation of Arm-optimized libraries. We present the optimizations implemented in each kernel and a detailed performance evaluation and comparison of their performance on four different HPC machines (i.e., A64FX, Graviton3, Intel Xeon Skylake Platinum, and AMD EPYC Rome). Overall, the experimental evaluation shows that Graviton3 outperforms other machines on average. Moreover, we observed that the performance of the A64FX is significantly constrained by its small memory hierarchy and latencies. Additionally, as proof of concept, we study the performance of a production-ready tool that exploits two of the ported and optimized genomic kernels.

Federated Learning Framework for Human Activity Recognition Using Smartphones

Human activity recognition plays a vital role in various applications like healthcare, sports, and smart environments. … With the increase in the use of smart wearables and sensors, the concern for privacy has increased drastically in recognizing the activities of individuals. The legacy techniques send the data from millions of users to a central server where the processing and model generation happens. This increases the load on the server. Federated Learning is a promising method to handle sensitive and personal data in an efficient manner as well as use distributed techniques to reduce challenges that are faced in legacy approaches. Edge devices like smartphones and embedded devices extract the raw data, preprocess it, and generate a personalized model. These models are then aggregated at a central server using federated learning methods and their aggregate model is sent back to clients. Long Short-Term Memory (LSTM) is used for model generation as it can effectively capture temporal dependencies in sequential data. In order to show a real-world scenario, an AWS EC2 instance is used as a server, and laptops and smartphones are used as clients. Our results demonstrate that such methods can be adopted and can be used on a much larger scale. The performance of the global and local models is comparable to those attained by conventional methods. The utilization of federated learning shows a significant step forward in the development of human activity recognition. This approach allows for collaborative learning across multiple devices while also ensuring the privacy of users is maintained.

Lessons Learned from Nanbar Health: Developing a Multi-System Software Suite for Predictive Algorithms and Remote Monitoring for People Living with Sickle Cell Disease

Background: Sickle cell disease (SCD) is a genetic disorder that alters the shape of red blood cells, leading to significant complications and requiring highly complex care. Efforts often focus on prevention and early diagnosis of pain to improve outcomes for those with SCD. Remote monitoring (RPM) and predictive algorithms has grown exponentially in the last decade as a support tool for managing chronic diseases like heart failure and diabetes. Implementing similar technology for SCD care could improve patient outcomes through early detection of complications and providing physicians with longitudinal data to support clinical decisions. Our team previously used mobile health (mHealth) as RPM in this population and developed a machine learning algorithm to predict pain scores with 86% accuracy. Here, we describe the development process for the Nanbar Health software alongside the lessons learned from its development. Objectives: Describe the methods necessary to build a comprehensive mHealth remote monitoring system (Nanbar Health) that facilitates SCD health management. Develop a set of lessons for other developers when building a remote monitoring software suite. Methods: Nanbar Health's development process has taken 2 years. First, our team reviewed feedback and “bug” reports from patients and providers of other mobile applications our team built at Duke University. Using the feedback collected an user-interface (UI) interface (mock-ups) was created for the Nanbar Health mobile app using Figma. These mock-ups were then shown to healthcare providers with expertise in caring for individuals with SCD for feedback. Upon approval of mock-ups, our team developed the mobile application using Swift and Apple HealthKit. The team also built a physician-facing dashboard using React, NodeJS, and Python. The backend software for this system utilized Docker containers mounted on AWS EC2 instances, a Parse server, MongoDB and PostgreSQL databases, and Nginx for deployment and load balancing. Next, the dashboard was reviewed by medical providers for feedback. The analytics portion of the dashboard processed data in Python backend scripts and interactive visualizations were built using the Plotly library in NodeJS. The analytics portion of the tool is currently being reviewed for feedback. Throughout the process of developing and piloting the Nanbar Health Software, our team identified several challenges. These challenges were collected into a comprehensive index named, “Lessons Learned”. Results: The lessons learned include: 1. Overall Gain user feedback and place yourself in the shoes of the user: At every step of your software's development ensure you are continuously getting feedback from your users. 2. Mobile Application Engagement is key to acquiring good data: Ensure your team uses a variety of methods methods to engage your users.No mobile app is one size fits all, but one app can customize to all: The development team should create re-usable, dynamic components wherever possible. 3. Provider Dashboard Testing to build: Developers should build a user interface that is easy to run and test as a “pseudo-user” while they are building it, making it easier for the developer to understand the user experience and build a user-friendly application.Building to test: Plan for beta testers from the beginning and ensure that they have access to continuously-updated and fake but realistic data. 4. Network Analysis Complex analytics are computationally demanding - Advanced analytical tools incur more complications when incorporated into a High-Usage SystemFor heavy data processing, go serverless - Isolating heavy data processing from the rest of the application saves tremendous time and money. 5. Research and Operations Don't cry over spilt code: Mistakes and bugs are part of the process when developing any software, and part of the research process.Lack of engagement is not a reason for disappointment, but rather an opportunity for growth and improvement: Our tool has continuously experienced changes and improvements to increase patient engagement. Conclusion: Our team was successfully able to create and deploy a SCD care-focused Nanbar Health. These tools show promise of improving care for people living with SCD through the usage of RPM and predictive models. Further research should be conducted to determine methods to increase user engagement, and maintain large predictive models in cloud-based systems.

Adaptive Load Balancing Using RR and ALB: Resource Provisioning in Cloud

Cloud Computing context, load balancing is an issue. With a rise in the number of cloud-based technology users and their need for a broad range of services utilizing resources successfully or effectively in a cloud environment is referred to as load balancing, has become a significant obstacle. Load balancing is crucial in storage systems to increase network capacity and speed up response times. The main goal is to present a new load-balancing mechanism that can balance incoming requests from users all over globally who are in different regions requesting data from remote data sources. This method will combine effective scheduling and cloud-based techniques. A dynamic load balancing method was developed to ensure that cloud environments have the ability to respond rapidly, in addition to running cloud resources efficiently and speeding up job processing times. Applications' incoming traffic is automatically split up across a number of targets, including Amazon EC2 instances, network addresses, and other entities by elastic load balancing. Elastic load balancing offers three distinct classifications of load balancer, and each one provides high availability, intelligent scalability, and robust security to guarantee the error-free functioning of your applications. Application load balancing and round robin are the two load balancing mechanisms in database cloud that are focus of this research study.

CAMPUS CLOUD: EMPOWERING UNIVERSITY MANAGEMENT WEB APPLICATION WITH CLOUD-HOSTED DOCKER TECHNOLOGY ON AWS

Abstract—In this article I have deployed a web-based applica- tion to Docker using Amazon web service (EC2 Instance). Elastic Cloud Compute service is used to the run the docker applications on scalable clusters. Amazon EC2 is fully managed and secured Container orchestration service that makes to deploy, manage and scale Dockerized web application. Docker containers on an EC2 instance to run a university management system website on the Amazon Web Services (AWS) cloud. For educational institutions, a university management system is a crucial tool that enables administrators to manage student records, course information, and other crucial data. The introduction to Docker, a containerization technology that makes it simple to build and manage web applications, follows a discussion of the advantages of hosting a university management system website on the cloud. The next section of the paper focuses on the AWS EC2 service and its capabilities, such as its capability for flexible computing and scalability. The paper also offers a step-by-step tutorial on how to use Docker to deploy a university administration system website on an EC2 instance, including instructions on how to set up the EC2 instance, install and configure Docker, and deploy the website code in a Docker container. Elastic Load Balancer and Auto Scaling, two AWS services, are used in the process of setting up security and scalability for the website. The benefits of using Docker to host a university administration system website on AWS EC2 are outlined in the paper’s conclusion, along with predictions for the development of cloud-based web application hosting. This research paper offers a thorough overview of the steps involved in hosting a university management system website using Docker on AWS EC2, making it a useful tool for educational institutions looking to improve their information management procedures and take advantage of cloud computing and Docker- Containerization.

СПОСОБИ З’ЄДНАННЯ З ВІРТУАЛЬНИМ СЕРВЕРОМ AWS, ЩО РОЗТАШОВАНИЙ В ПРИВАТНІЙ ПІДМЕРЕЖІ

In today's world of total digitization cyber security and safe work with data in cyberspace are the most important questions. Especially this is actual for Ukraine, where the number and power of cyberattacks has increased several times over the last year. Businesses that work with private customer data become especially vulnerable. Of course, the ideal option is to place such data on servers that don’t have Internet access. But according to the global trend of moving to the cloud, it is inevitable for private data as well. And so, there is a question of protecting private data in the cloud. To this end, cloud service providers offer services to create private subnets without Internet access. Therefore, the question of how to securely access data in such subnets become actual. One of the leader’s vendors in cloud servicing is Amazon with its Web Services. Amazon offers a Virtual Private Cloud service for setting up a virtual network. The article deals with the analysis of configuration features at the stage of creation of subnets with and without Internet access. The method of connection to a virtual server, located in a private subnet, using the Secure Shell network protocol was analyzed. However, this method has a number of disadvantages. It requires to launch an additional server and its administration. The method also has quite complex settings of the network and requires managing keys. Therefore, another method of connection to private EC2 instance was proposed. The method requires Amazon Systems Manager service, which provides secure access to data without creating additional server, is cost-effective and convenient. At the same time, all connections take place over a secure channel between the Systems Manager agent and the Amazon data center. Main setting’s features for the proposed method were considered.

SFCaaS: Service function chains as a service in NFV environments

With the growing deployment of emergent technologies like software-defined networking, network services are expected to be revolutionized. In this paper, we investigate offering Service Function Chains as a Service (SFCaaS) in NFV environments. We describe the potential business model to offer such a service and then we address the service function chain provisioning and resource allocation problem. As the chain is deployed thanks to virtual machines (i.e., instances) and links, we conduct first a detailed study of the costs of Amazon EC2 instances with respect to their location, size, type and performance. Afterwards, we address the resource allocation problem for service function chains from the SFC provider's perspective. We formulate the problem as an integer linear program aiming at reducing operational costs of the service function chains (i.e., costs of virtual machine instances and links, and synchronization among the instances). To address large-scale instances of the problem, we also propose a new heuristic algorithm to reduce operational costs taking into account the conducted study of the costs of Amazon EC2 instances. We show through extensive simulations that the proposed heuristic significantly reduces operational costs compared to a baseline algorithm inspired by the existing literature.

Secure Data Mobility in Cloud Computing for e-Governance Application

Due to the rapid technological advancement, traditional e-government systems are getting obsolete because of their inherent limitation of interoperability and accessibility to the highly secured and flexible e-governance services. Migration of such systems into highly secured cloud governance architecture will be a long-term viable solution. However, the adoption of distributed cloud computing has created operational and security challenges. This research work aims to bridge the gap between traditional and cloud-based e-Government systems in terms of data security based on confidentiality, interoperability, and mobility of data among distributed databases of cloud computing environments. In this work, we have created two organization databases by the use of AWS EC2 instances and classified the data based on the Risk Impact Level (RIL) of data by the use of the Metadata Attribute Value (MAV) function. To enhance further security on classified data, we take appropriate security action based on the sensitivity of the data. For the analysis purpose, we implemented different security algorithms, i.e. AES, DES, and RSA in the mobility of data between two distributed cloud databases. We measured the encryption and decryption time along with the file size of data before and after classification. AES performed better while considering the encryption time and file size, but the overall performance of RSA was better for smaller file sizes. Finally, the performance of the data mobility between two distributed clouds databases was analyzed while maintaining the sensitivity level of the data.

Research on Machine Learning Optimization Algorithm of CNN for FPGA Architecture

In order to meet the application requirements of deep learning detection and recognition algorithm for Field Programmable Gate Array (FPGA) computing structure, this paper proposes a machine learning optimization algorithm of convolutional neural network (CNN) based on pruning and Int8 quantization. By analyzing the machine learning intelligent recognition network, the Yolo v3 target detection network is selected to verify the optimization algorithm. Aiming at the Yolo v3 network with balanced speed and precision, the optimization algorithm is studied in detail. Based on Amazon Cloud FPGA EC2 instance platform and ZCU104 FPGA hardware platform, machine learning optimization algorithms are used to accelerate the application of Yolo v3 network detection and recognition. Satisfactory results are obtained on both FPGA computing platforms. The experimental results show that the machine learning optimization technology can improve the inference speed of neural network based on both FPGA platforms. The network structure optimized by pruning and Int8 quantization algorithms can achieve high acceleration speed, with very limited accuracy loss.

Infrastructure Optimization in Kubernetes Cluster

Kubernetes is a compact, extensible, open-source stage for overseeing containerized responsibilities and administrations, that works with both decisive setup and robotization. Kubernetes is like VMs, however having loosened up isolation properties to share the Operating System (OS) among the applications. The container conversely with VM, has its own document framework, a portion of Central Processing Unit(CPU), memory, process space, and much more. Kubernetes cluster is a bunch of node machines for running containerized applications. Each cluster contains a control plane and at least one node. Infrastructure Optimization is the process of analyzing and arranging the portion of cloud resources that power applications and workloads to augment the presentation and limit squander due to over-provisioning. In the paper, a “Movie Review System” web application is designed using GoLang for backend components and HTML, CSS, and JS for frontend components. Using AWS, an EC2 instance is created and the web application is deployed onto EC2 and hosted in the instance server. The web application is also deployed on Kubernetes locally using the MiniKube tool. A performance analysis is performed for both the deployments on considering common performance metrics for both AWS EC2 / Virtual Machine (VM) and Kubernetes.

SDN-Based Network Slicing Mechanism for a Scalable 4G/5G Core Network: A Kubernetes Approach.

Managing the large volumes of IoT and M2M traffic requires the evaluation of the scalability and reliability for all the components in the end-to-end system. This includes connectivity, mobile network functions, and application or services receiving and processing the data from end devices. Firstly, this paper discusses the design of a containerized IoT and M2M application and the mechanisms for delivering automated scalability and high availability when deploying it in: (1) the edge using balenaCloud; (2) the Amazon Web Services cloud with EC2 instances; and (3) the dedicated Amazon Web Services IoT service. The experiments showed that there are no significant differences between edge and cloud deployments regarding resource consumption. Secondly, the solutions for scaling the 4G/5G network functions and mobile backhaul that provide the connectivity between devices and IoT/M2M applications are analyzed. In this case, the scalability and high availability of the 4G/5G components are provided by Kubernetes. The experiments showed that our proposed scaling algorithm for network slicing managed with SDN guarantees the necessary radio and network resources for end-to-end high availability.

Highly scalable and load balanced web server on AWS cloud

Most of the industries maintaining their web server on their on-premise, which leads them to have high maintenance, more workload, increases cost. To overcome this we can use cloud technologies for building a web server which gives better performance. Thus we use AWS cloud technologies. In this research paper, a web server is built using the cloud formation technique CFT. Using this technique we can able to build our infrastructure using YAML code. To achieve these major cloud resources are used such as Ec2 Instance, Application load balancer for balancing the load given by users, auto-scaling group. We will configure these cloud resources as YAML and implement them using the cloud formation technique. So that the industries can have their Infrastructure as Code.

Power efficient virtual machine placement in cloud data centers with a discrete and chaotic hybrid optimization algorithm

Cloud computing is a new computation technology that provides services to consumers and businesses. The main idea of Cloud computing is to present software and hardware services through the Internet to the users and organizations at all levels. In Cloud computing, the users pay for the services, which means a usage-based payment system is used in this technology. Using virtualization technology in computation resources enables the appropriate utilization of resources in cloud computing. One of the most significant challenging issues in virtualization technology is the placement of optimal virtual machines on physical machines in cloud data centers. The placement of virtual machines comprises a process wherein virtual machines are mapped onto physical machines in cloud data centers. Optimal deployment leads to the reduction in power consumption, optimal use of resources, traffic reduction in data centers, costs reduction, and efficiency enhancement of the data center in the cloud. The present article proposed a new approach using a combination of the Sine–Cosine Algorithm and Salp Swarm Algorithm as discrete multi-objective and chaotic functions for optimal virtual machine placement. The first goal of the proposed algorithm was to reduce the power consumption in cloud data centers by condensing the number of active physical machines. The second goal was to reduce the waste of resources and manage it by optimally virtual machine placement on physical machines in cloud data centers. The third objective was to minimize and reduce Service Level Agreement among the active physical machines in cloud data centers. The proposed method prevent the increase in the migration of virtual machines onto physical machines. Ultimately, the results obtained from the proposed algorithm were compared with those of previous akin algorithms in the literature, including First Fit, Virtual Machine Placement Ant Colony System, and Modified Best Fit Decreasing. The proposed scheme is tested using Amazon EC2 Instances and the result indicated that the proposed algorithm performs better than the existing algorithms for various performance metrics.

Identification and characterization of tweets related to the 2015 Indiana HIV outbreak: A retrospective infoveillance study.

From late 2014 through 2015, Scott County, Indiana faced an HIV outbreak triggered by opioid abuse and transition to injection drug use. Investigating the origins, risk factors, and responses related to this outbreak is critical to inform future surveillance, interventions, and policymaking. In response, this retrospective infoveillance study identifies and characterizes user-generated messages related to opioid abuse, heroin injection drug use, and HIV status using natural language processing (NLP) among Twitter users in Indiana during the period of this HIV outbreak. Our study consisted of two phases: data collection and processing, and data analysis. We collected Indiana geolocated tweets from the public Twitter API using Amazon Web Services EC2 instances filtered for geocoded messages in the immediate pre and post period of the outbreak. In the data analysis phase we applied an unsupervised machine learning approach using NLP called the Biterm Topic Model (BTM) to identify tweets related to opioid, heroin/injection, and HIV behavior and then examined these messages for HIV risk-related topics that could be associated with the outbreak. More than 10 million geocoded tweets occurring in Indiana during the immediate pre and post period of the outbreak were collected for analysis. Using BTM, we identified 1350 tweets thought to be relevant to the outbreak and then confirmed 358 tweets using human annotation. The most prevalent themes identified were tweets related to self-reported abuse of illicit and prescription drugs, opioid use disorder, self-reported HIV status, and public sentiment regarding the outbreak. Geospatial analysis found that these messages clustered in population dense areas outside of the outbreak, including Indianapolis and neighboring Clark County. This infoveillance study characterized the social media conversations of communities in Indiana in the pre and post period of the 2015 HIV outbreak. Behavioral themes detected reflect discussion about risk factors related to HIV transmission stemming from opioid and heroin abuse for priority populations, and also help identify community attitudes that could have motivated or detracted the use of HIV prevention methods, along with helping identify factors that can impede access to prevention services. Infoveillance approaches, such as the analysis conducted in this study, represent a possibly strategy to detect "signal" of the emergence of risk factors associated with an outbreak though may be limited in their scope and generalizability. Our results, in conjunction with other forms of public health surveillance, can leverage the growing ubiquity of social media platforms to better detect opioid-related HIV risk knowledge, attitudes and behavior, as well as inform future prevention efforts.

Creating a Custom Virtual Private Cloud and Launch an Elastic Compute Cloud (EC2) Instance in Your Virtual Private Cloud

Cloud Computing is regarded to as putting away and getting to data over the web. The hard disk of your PC doesn’t hold this data. In Cloud computing, you can get to information from a remote server. Amazon Web Services (AWS) enables adaptable, solid, versatile, simple to-utilize and practical Cloud computing arrangements. AWS is an extensive, simple to utilize processing stage offered Amazon. A virtual private cloud (VPC) is devoted to the AWS account which is in the AWS cloud that acts coherently detached with different virtual systems. Amazon EC2 is a secure web administration which allows register with the modifiable limit in the cloud .In this work, we are giving route subtleties of Creating a custom VPC and dispatch an EC2 Instance in your VPC.

Stanza: Layer Separation for Distributed Training in Deep Learning

The parameter server architecture is prevalently used for distributed deep learning. Each worker machine in a such system trains the complete model, which leads to a large amount of network data transfer between workers and servers. We empirically observe that the data transfer has a major impact on training time. We present a new distributed training system called Stanza to tackle this problem. Stanza exploits the fact that in many models such as convolution neural networks, most data exchange is attributed to the fully connected layers, while most computation is carried out in convolutional layers. Thus, we propose layer separation in distributed training: most nodes of the cluster train only the convolutional layers, while the rest train the fully connected layers. Gradients and parameters of the fully connected layers no longer need to be exchanged across the entire cluster, thereby substantially reducing the data transfer volume. We implement Stanza on PyTorch and evaluate its performance on Azure and EC2. Results show that Stanza accelerates training significantly over current parameter server systems: on EC2 instances with Tesla V100 GPU and 10Gb bandwidth for example, Stanza is 1.34x-13.9x faster for common deep learning models.

Cloud Computing and Local Chip-Based Dynamic Economic Dispatch for Microgrids

Dynamic economic dispatch in microgrids is usually realized in a centralized energy management system (EMS). However, centralized systems are subject to single-point failure problems. In rural areas and islands, a failed EMS cannot be recovered in a timely manner due to lack of technical support. In this paper, we propose a cloud and edge computing-based framework to realize dynamic economic dispatch, which is conducted on a local Digital Signal Processor (DSP) chip and a remote cloud computing platform (CCP). Based on the multi-parametric programming algorithm, the dispatch process is divided into two parts: offline calculation and real-time decision making. These calculation tasks can be conducted on the remote cloud computing platform and the DSP chip of the inverter in an arbitrary renewable generator, respectively. The tests are carried out on the Amazon Web Service (AWS) EC2 instance and an autonomous microgrid with DSP TMS320F28335 chip. The results show that the proposed method can obtain the same results as the conventional method, with significantly improved reliability.

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.

An Optimal Recovery Approach for Liberation Codes in Distributed Storage Systems

To reduce the storage cost, distributed storage systems are gradually using erasure codes to ensure data reliability. Liberation codes, which satisfy the maximum distance separable (MDS) property and provide optimal modification overhead, are a class of popular two fault tolerant erasure codes. However, erasure codes need to read from surviving nodes and transfer across the network large amounts of data when recovering from single node failures. Existing single node failure recovery approaches for Liberation codes are either time-consuming or suboptimal. In this article, firstly, we prove the minimum number of symbols required to recover one failed node for a Liberation coded system. Then we derive the conditions that optimal recovery solutions need to satisfy. Finally, we propose an algorithm, called Disk Read Optimal Recovery (DROR), which can determine an optimal recovery solution in linear time and recover the failed node reading the minimum amount of data. We have implemented DROR in a real-world storage system Ceph and evaluated DROR on a cluster of Amazon EC2 instances. We show that DROR reduces the reconstruction time by up to 23.6% compared to that of the recovery approach in Ceph.