Computing Data Centers Research Articles

Orchestration of Ultra-Dense 5G Networks

The six articles in this special section address ultra-dense 5G mobile communications networks. The articles present the research and industrial communities with the opportunity to explore new research findings on ultra-dense networks employing network function virtualization (NFV)/software defined networks (SDN) in their cloud-based architecture. The focus is on the use of virtual management and orchestration features to drive the various network resources in response to the load changes at the RAN and distributing traffi c between various cloud entities. The NFV transformation is gaining incredible momentum within mobile operators as one of the prominent solutions to improve the resource allocation and system scalability in future fi fth generation (5G) networks. The process of migrating proprietary hardware solutions to the NFV domain creates a complete new cycle of dynamic service management that can adapt in response to users’ behavior and resource availability. Virtual appliances, denoted as virtual machines (VMs), are normally dependent on the underlying cloud hardware through a thin software layer that abstracts virtual resources to NFV applications. 5G networks will adopt a distributed cloud architecture that supports local processing of user requests within mobile edge computing (MEC) data centers. In addition, the 5G core network (5GC) will be deployed as a virtual composition that can instantiate interfaces to various radio access network (RAN) technologies over the cloud.

Joint Minimization of the Energy Costs From Computing, Data Transmission, and Migrations in Cloud Data Centers

We propose a novel model, called joint computing, data transmission and migration energy costs (JCDME), for the allocation of virtual elements (VEs), with the goal of minimizing the energy consumption in a software-defined cloud data center (SDDC). More in detail, we model the energy consumption by considering the computing costs of the VEs on the physical servers, the costs for migrating VEs across the servers, and the costs for transferring data between VEs. In addition, JCDME introduces a weight parameter to avoid an excessive number of VE migrations. Specifically, we propose three different strategies to solve the JCDME problem with an automatic and adaptive computation of the weight parameter for the VEs migration costs. We then evaluate the considered strategies over a set of scenarios, ranging from a small sized SDDC up to a medium-sized SDDC composed of hundreds of VEs and hundreds of servers. Our results demonstrate that JCDME is able to save up to an additional 7% of energy with respect to previous energy-aware algorithms, without a substantial increase in the solution complexity.

Drawing the Cloud Map: Virtual Network Provisioning in Distributed Cloud Computing Data Centers

Efficient virtualization methodologies constitute the core of cloud computing data center implementation. Clients are attracted to the cloud model by its ability to scale the resources dynamically and the flexibility in payment options that it offers. However, performance hiccups may push them to go back to the buy-and-maintain model. Virtualization plays a key role in the synchronous management of the thousands of servers along with clients' data living on them. To achieve seamless virtualization, cloud providers require a system that performs the function of virtual network provisioning. This includes receiving the cloud client requests and allocating their computational and network resources in a way that guarantees the quality-of-service conditions for clients while maximizing the data center resource utilization and providers' revenue. We introduce a comprehensive system to solve the problem of virtual network mapping for a set of connection requests sent by cloud clients. Connections are collected in time intervals called windows. Consequently, node and link provisioning is performed. Different window size selection schemes are introduced and evaluated. Three schemes to prioritize connections are used, and their effect is assessed. Moreover, a technique dealing with connections spanning over more than a window is introduced. The proposed algorithm is compared with previous work well known in the literature. Simulation results show that the dynamic window size algorithm achieves cloud service providers' objectives in terms of generated revenue, served-connection ratio, resource utilization, and computational overhead. In addition, experimental results show that handling spanning connections independently improves the performance of the system.

An Integer Linear Programming model and Adaptive Genetic Algorithm approach to minimize energy consumption of Cloud computing data centers

Cloud computing infrastructures are designed to support the accessibility and availability of various services to consumers over the Internet. Data centers hosting Cloud applications consume massive amount of power, contributing to high carbon footprints to the environment. Hence, solutions are needed to minimize the energy consumption. This paper focuses on the development of a dynamic task scheduling algorithm by proposing an Integer Linear Programming (ILP) model that minimizes the energy consumption in a Cloud data center. Furthermore, an Adaptive Genetic Algorithm (GA) is proposed to reflect the dynamic nature of the Cloud environment and to provide a near optimal scheduling solution that minimizes the energy consumption. The proposed adaptive GA is validated by simulating the Cloud infrastructure and conducting a set of performance and quality evaluation study in this environment. The results demonstrate that the proposed solution offers performance gains with regards to response time and in reducing energy consumption.

A Descriptive Study towards Green Computing Practice Application for Data Centers in IT Based Industries

The progressive upsurge in demand for processing and computing power has led to a subsequent upsurge in data center carbon emissions, cost incurred, unethical waste management, depletion of natural resources and high energy utilization. This raises the issue of the sustainability attainment in data centers of Information Technology (IT) based industries. Green computing practice can be applied to facilitate sustainability attainment as IT based industries utilizes data centers to provide services to staffs, practitioners and end users. But it is a known fact that enterprise servers utilize huge quantity of energy and incur other expenditures in cooling operations and it is difficult to address the needs of accuracy and efficiency in data centers while yet encouraging a greener application practice alongside cost reduction. Thus this research study focus on the practice application of Green computing in data centers which houses servers and as such presents the Green computing life cycle strategies and best practices to be practiced for better management in data centers in IT based industries. Data was collected through questionnaire from 133 respondents in industries that currently operate their in-house data centers. The analysed data was used to verify the Green computing life cycle strategies presented in this study. Findings from the data shows that each of the life cycles strategies is significant in assisting IT based industries apply Green computing practices in their data centers. This study would be of interest to knowledge and data management practitioners as well as environmental manager and academicians in deploying Green data centers in their organizations.

Modelling and Analysis of A Novel Deadline-Aware Scheduling Scheme for Cloud Computing Data Centers

User request (UR) service scheduling is a process that significantly impacts the performance of a cloud data center. This is especially true since essential quality-of-service (QoS) performance metrics such as the UR blocking probability as well as the data center's response time are tightly coupled to such a process. This paper revolves around the proposal of a novel Deadline-Aware UR Scheduling Scheme (DASS) that has the objective of improving the data center's QoS performance in term of the above-mentioned metrics. A minority of existing work in the literature targets the formulation of mathematical models for the purpose of characterizing a cloud data center's performance. As a contribution to covering this gap, this paper presents an analytical model, which is developed for the purpose of capturing the system's dynamics and evaluating its performance when operating under DASS. The model's results’ accuracy are verified through simulations. Also, the performance of the data center achieved under DASS is compared to its counterpart achieved under the more generic First-In-First-Out (FIFO) scheme. The reported results indicate that DASS outperforms FIFO by $11$ to $58$ percent in terms of the blocking probability and by $82$ to $89$ percent in terms of the system's response time.

A power saver scheduling algorithm using DVFS and DNS techniques in cloud computing data centres

Cloud computing is a fascinating and profitable area in modern distributed computing. Aside from providing millions of users the means to use offered services through their own computers, terminals, and mobile devices, cloud computing presents an environment with low cost, simple user interface, and low power consumption by employing server virtualisation in its offered services (e.g., Infrastructure as a Service). The pool of virtual machines found in a cloud computing data centre (DC) must run through an efficient task scheduling algorithm to achieve resource utilisation and good quality of service, thus ensuring the positive effect of low energy consumption in the cloud computing environment. In this paper, we present an energy-efficient scheduling algorithm for a cloud computing DC using the dynamic voltage frequency scaling technique. The proposed scheduling algorithm can efficiently reduce the energy consumption for executing jobs by increasing resource utilisation. GreenCloud simulator is used to simulate our algorithm. Experimental results show that, compared with other algorithms, our algorithm can increase server utilisation, reduce energy consumption, and reduce execution time.

A power saver scheduling algorithm using DVFS and DNS techniques in cloud computing data centres

Cloud computing is a fascinating and profitable area in modern distributed computing. Aside from providing millions of users the means to use offered services through their own computers, terminals, and mobile devices, cloud computing presents an environment with low cost, simple user interface, and low power consumption by employing server virtualisation in its offered services (e.g., Infrastructure as a Service). The pool of virtual machines found in a cloud computing data centre (DC) must run through an efficient task scheduling algorithm to achieve resource utilisation and good quality of service, thus ensuring the positive effect of low energy consumption in the cloud computing environment. In this paper, we present an energy-efficient scheduling algorithm for a cloud computing DC using the dynamic voltage frequency scaling technique. The proposed scheduling algorithm can efficiently reduce the energy consumption for executing jobs by increasing resource utilisation. GreenCloud simulator is used to simulate our algorithm. Experimental results show that, compared with other algorithms, our algorithm can increase server utilisation, reduce energy consumption, and reduce execution time.

MooreCube: A Scalable and Flexible Architecture for Cloud Computing Data Centers on Multi-port Servers

Networks provide the infrastructure of cloud computing data centers. Servers in data centers grow to respond to the growing user demands. Data center scalability and flexibility in the use of multi-port servers is a challenging issue. Heretofore, no method has been considered for scalability and flexibility issues. This paper proposes a new architecture called MooreCube that can increase network scalability and decrease network diameter as well as increase flexibility. MooreCube is a scalable and flexible architecture that each multi-port server directly connected to other servers via bi-directional links, without using any switch. Furthermore, MooreCube is a recursively defined architecture that uses Moore graph as Building Block (BB) structure and uses the hierarchical structure to meet high scalability. The paper proposes a multipath routing to increase fault tolerance and decrease links burden. MooreCube architecture compared with other switchless architectures that use reserved ports to increase scalability. The simulation results show MooreCube increase scalability and flexibility along with decrease the diameter of the network.

An efficient and improved multi-objective optimized replication management with dynamic and cost aware strategies in cloud computing data center

Recent technology which focus in cloud computing with ICT based service providers for better challenges in the field of online services. This provides the computing world as an utility based scenario of sharing the given resources of different data centeres with enormous opportunities. Though replicas of a data file has increased, the performance and availability of data also increases. In this paper the different aspects on dynamic, cost-aware with the data replication method through optimization is proposed which identifies the less amount of information of data replication which is required to make sure that the data availability increases with the increase in the replication process. The multi objective optimization strategy for the cost of replication from higher-cost data centers to lower-cost data centers being implemented by the concept of Improved knapsack algorithm and considering the availability and load balancing in the replication process. Data can be managed effectively by file replication, which reduces effectively the file service time and access latency, which increases the file availability and the system get improved through load balancing. An efficient and improved multi-objective optimized replication management (EIMORM) can solve these optimal solutions by balancing among these optimization objectives. Some experiments clearly shows that the EIMORM is much more energy efficiency and the performance in the replication system of the Hadoop Distributed File System (HDFS) with the multi-objective evolutionary stated algorithm with the performance, load balancing in cloud storage clustors. Now considering the energy efficiency also considering the bandwidth of system. Hence the results taken through simulator can support the energy efficiency as the guidance in the areas of data replication.

Resource Isolation Method for Program’S Performance on CMP

Data center and cloud computing are more popular, which make more benefits for customers and the providers. However, in data center or clusters, commonly there is more than one program running on one server, but programs may interference with each other. The interference may take a little effect, however, the interference may cause serious drop down of performance. In order to avoid the performance interference problem, the mechanism of isolate resource for different programs is a better choice. In this paper we propose a light cost resource isolation method to improve program’s performance. This method uses Cgroups to set the dedicated CPU and memory resource for a program, aiming at to guarantee the program’s performance. There are three engines to realize this method: Program Monitor Engine top program’s resource usage of CPU and memory and transfer the information to Resource Assignment Engine; Resource Assignment Engine calculates the size of CPU and memory resource should be applied for the program; Cgroups Control Engine divide resource by Linux tool Cgroups, and drag program in control group for execution. The experiment result show that making use of the resource isolation method proposed by our paper, program’s performance can be improved.

The role of dedicated data computing centers in the age of cloud computing

Brookhaven National Laboratory (BNL) anticipates significant growth in scientific programs with large computing and data storage needs in the near future and has recently reorganized support for scientific computing to meet these needs. A key component is the enhanced role of the RHIC-ATLAS Computing Facility (RACF) in support of high-throughput and high-performance computing (HTC and HPC) at BNL. This presentation discusses the evolving role of the RACF at BNL, in light of its growing portfolio of responsibilities and its increasing integration with cloud (academic and for-profit) computing activities. We also discuss BNL’s plan to build a new computing center to support the new responsibilities of the RACF and present a summary of the cost benefit analysis done, including the types of computing activities that benefit most from a local data center vs. cloud computing. This analysis is partly based on an updated cost comparison of Amazon EC2 computing services and the RACF, which was originally conducted in 2012.

Energy-efficient framework for virtual machine consolidation in cloud data centers

With the advent of the era of cloud computing, the high energy consumption of cloud computing data centers has become a prominent problem, and how to reduce the energy consumption of cloud computing data center and improve the efficiency of data center has become the research focus of researchers all the world. In a cloud environment, virtual machine consolidation (VMC) is an effective strategy that can improve the energy efficiency. However, at the same time, in the process of virtual machine consolidation, we need to deal with the tradeoff between energy consumption and excellent service performance to meet service level agreement (SLA). In this paper, we propose a new virtual machine consolidation framework for achieving better energy efficiency-Improved Underloaded Decision (IUD) algorithm and Minimum Average Utilization Difference (MAUD) algorithm. Finally, based on real workload data on PlanetLab, experiments have been done with the cloud simulation platform CloudSim. The experimental result shows that the proposed algorithm can reduce the energy consumption and SLA violation of data centers compared with existing algorithms, improving the energy efficiency of data centers.

Performance and Energy Efficiency Metrics for Communication Systems of Cloud Computing Data Centers

Cloud computing has become a de facto approach for service provisioning over the Internet. It operates relying on a pool of shared computing resources available on demand and usually hosted in data centers. Assessing performance and energy efficiency of data centers becomes fundamental. Industries use a number of metrics to assess efficiency and energy consumption of cloud computing systems, focusing mainly on the efficiency of IT equipment, cooling and power distribution systems. However, none of the existing metrics is precise enough to distinguish and analyze the performance of data center communication systems from IT equipment. This paper proposes a framework of new metrics able to assess performance and energy efficiency of cloud computing communication systems, processes and protocols. The proposed metrics have been evaluated for the most common data center architectures including fat tree three-tier, BCube, DCell and Hypercube.

Taming the Costs of Trustworthy Provenance through Policy Reduction

Provenance is an increasingly important tool for understanding and even actively preventing system intrusion, but the excessive storage burden imposed by automatic provenance collection threatens to undermine its value in practice. This situation is made worse by the fact that the majority of this metadata is unlikely to be of interest to an administrator, instead describing system noise or other background activities that are not germane to the forensic investigation. To date, storing data provenance in perpetuity was a necessary concession in even the most advanced provenance tracking systems in order to ensure the completeness of the provenance record for future analyses. In this work, we overcome this obstacle by proposing a policy-based approach to provenance filtering , leveraging the confinement properties provided by Mandatory Access Control (MAC) systems in order to identify and isolate subdomains of system activity for which to collect provenance. We introduce the notion of minimal completeness for provenance graphs, and design and implement a system that provides this property by exclusively collecting provenance for the trusted computing base of a target application. In evaluation, we discover that, while the efficacy of our approach is domain dependent, storage costs can be reduced by as much as 89% in critical scenarios such as provenance tracking in cloud computing data centers. To the best of our knowledge, this is the first policy-based provenance monitor to appear in the literature.

Solid-State Drives: Memory Driven Design Methodologies for Optimal Performance

Solid-state drives (SSDs) faced an astonishing development in the last few years, becoming the cornerstone to new paradigms and markets of the information technology, such as cloud computing and big data centers. So far, the SSD design approach has focused on the optimization of the Flash translation layer, the firmware devoted to fulfill the compatibility with traditional hard-disk drives. For hyperscaled SSDs this strategy is no longer valid since their performance and reliability are strictly linked to that of the NAND Flash memories that constitute the storage medium, in particular when the multilevel cell paradigm is considered. For this reason, the design flow must follow a bottom-up approach that, starting from an accurate knowledge of the time and use dependent reliability of the NAND Flash memories, selects the most appropriate error correction strategy to extend the SSD lifetime while reducing its performance degradation. Then, the design flow moves to that of the SSD controller and of the interface toward the host where the application is running. This paper will thoroughly discuss this bottom-up approach, and finally, it will show how it is possible to leverage new approaches, such as the software-defined storage system that, by exploiting a hardware/software codesign of the SSD controller architecture and of the host application, will be able to revolutionize the traditional computer/memory interaction.

MEC towards 5G: A Survey of Concepts, Use Cases, Location Tradeoffs

In recent years, there has been a new trend to push everything to a centralized cloud, triggered by virtualization and pushed by the need to reduce costs and increase suppleness. In the process, mobile operators and industry players forgot how prominent the location of the functionality is to performance, optimal use of network resources and user experience. As they progressively grasp the influence of location in wireless networks and specifically in virtualized networks they start to look for efficient ways to deploy network taking in consideration those metrics. On the one hand Distributed Cloud RAN (DC-RAN) which consists in deploying Base Band Units (BBUs) in distributed way instead of pooling the units at centralized data center and Multiple-access Edge Computing (MEC), Previously known as Mobile Edge Computing which consists in mixing the IT & Telco domains and bringing their capabilities within the close proximity of mobile subscribers to better serve them are gaining acceptance in mobile networks. On the other hand, Software Defined Networking (SDN) and Network Function Virtualization (NFV), two promising concepts are expected to take mobile networks to a high level of agility. While SDN is based on the separation of control and data planes, NFV separates software from hardware enabling flexible network deployment and dynamic operation. Virtualization, originally used as the support for shifting to the centralized cloud, is even more basic in permitting hybrid models because it offers service providers the chance to choose a location, hardware, and software separately to optimize end-to-end network performance and Quality of Experience (QoE). In this perspective, NFV, SDN, DC-RAN, and MEC in distinct but in complementary ways man up the necessity to put processing and storage where it’s suitable to preserve a healthful harmony between what lingers centralized and what have to be distributed to the edge based on parameters such as applications, traffic type, and network conditions. In this paper, we will present MEC use cases that have gained an attraction to date and we will shed the light on the importance of the edge location and criteria to take in consideration while deploying MEC within network along with the MEC location tradeoffs.

Understanding Big Data Analytics Workloads on Modern Processors

Big data analytics workloads are very significant ones in modern data centers, and it is more and more important to characterize their representative workloads and understand their behaviors so as to improve the performance of data center computer systems. In this paper, we embark on a comprehensive study to understand the impacts and performance implications of the big data analytics workloads on the systems equipped with modern superscalar out-of-order processors. After investigating three most important application domains in Internet services in terms of page views and daily visitors, we choose 11 representative data analytics workloads and characterize their micro-architectural behaviors by using hardware performance counters. Our study reveals that the big data analytics workloads share many inherent characteristics, which place them in a different class from the traditional workloads and the scale-out services. To further understand the characteristics of big data analytics workloads, we perform correlation analysis to identify the most key factors that affect cycles per instruction (CPI). Also, we reveal that the increasing complexity of the big data software stacks will put higher pressures on the modern processor pipelines.

A survey of memory deduplication approaches for intelligent urban computing

Limited memory size is considered as a major bottleneck in data centers for intelligent urban computing. It is shown that there exist a large number of duplicated pages when various processes working with big data are hosted in data centers. Memory deduplication aims to automatically eliminate duplicate data in memory. It is an efficient technique to reduce memory requirement. In memory deduplication, pages with same content are detected and merged into a single copy. Recently, several system-level techniques have been proposed to address this issue, in which content-based page sharing (CBPS) is most widely used, since CBPS can be performed transparently in the hypervisor layer without any modification to guest operating systems of data center. In this paper, we survey several techniques for memory deduplication. We also classify these techniques based on their characteristics to highlight their similarities and differences. The aim of this paper is to provide insights to researchers into working of memory deduplication techniques and also to motivate them to propose better intelligent urban computing systems.

Guest Editorial: Special Issue on Algorithms and Computational Models for Sustainable Computing in Cloud and Data Centers

The papers in this special section focus on the development of algorithms and computational models for sustainable computing in cloud and data centers. These papers bring together high quality contributions covering the topics of energy efficiency cloud computing, advances in sustainable cloud and data centers, energy aware applications at various levels of the computational and data storage processes, and emerging trends in sustainable computing. New research findings, novel approaches, and algorithms are presented for energy-efficient Cloud computing, energy-aware memory management solutions for data centers and processing of large data sets, energy aware scheduling approaches for data centers and processing of large data sets, energy consumption computational models, analytics and performance evaluations, and applications as well as emerging trends and applications.