Resource Fragmentation Research Articles

Intra-Balance Virtual Machine Placement for Effective Reduction in Energy Consumption and SLA Violation

Cloud computing has emerged as one of the most important technological revolutions globally. However, the rapid growth of cloud computing has imposed a massive financial burden and resulted in environmental side effects due to excessive energy consumption. The high power consumption is not only attributed to the size of data centers but also to the ineptitude of resource usage. Most of the extant research has focused on reducing power consumption by an aggressive VM consolidation, which leads to the violation of the service level agreement (SLA). Furthermore, the unbalanced resource consumption exacerbates the unavailable wasted resources that are referred to as unavailable resource fragmentation. In this paper, we propose the use of a balanced resources consumption algorithm called BRC-IBMMT in order to enhance the efficiency of resource consumption while achieving an acceptable balance between conflicting correlation objectives of power consumption as well as SLA violation. The extensive simulation results of different types of workload validate and lend credence to the significance of the proposed method in reducing power consumption and SLA violation of the cloud data center.

A dynamic cluster job scheduling optimisation algorithm based on data irreversibility in sensor cloud

The optimisation algorithm based on irreversible data for the job scheduling of dynamic cluster is crucial to the improvement of the cluster rendering throughput and the cluster rendering efficiency. However, the dispatching imbalance of the cluster rendering task on the massive number of cluster rendering nodes will prolong the waiting time for the completion of job. Therefore, we propose dynamic cluster job scheduling optimisation algorithm based on data irreversibility (DCJS_DI). Firstly, we analyse the job scheduling target. Then, we exploit the frame independence with the clustering rendering and further establish the job scheduling model for the irreversible data dynamic cluster. Next, we elaborate on the job scheduling process of the irreversible data dynamic cluster and the dispatching process of the cluster rendering task. Finally, we investigate via simulation results the impacts of the job hunger and the resource fragmentation issue of the traditional job scheduling strategies on the system performance, the impacts of the multi-progress and multi-threading cluster rendering on the job completion time and the resource efficiency of the irreversible data dynamic cluster. We further study the extension of the cluster computation capability and the reliability issue of the cloud service.

A dynamic cluster job scheduling optimisation algorithm based on data irreversibility in sensor cloud

The optimisation algorithm based on irreversible data for the job scheduling of dynamic cluster is crucial to the improvement of the cluster rendering throughput and the cluster rendering efficiency. However, the dispatching imbalance of the cluster rendering task on the massive number of cluster rendering nodes will prolong the waiting time for the completion of job. Therefore, we propose dynamic cluster job scheduling optimisation algorithm based on data irreversibility (DCJS_DI). Firstly, we analyse the job scheduling target. Then, we exploit the frame independence with the clustering rendering and further establish the job scheduling model for the irreversible data dynamic cluster. Next, we elaborate on the job scheduling process of the irreversible data dynamic cluster and the dispatching process of the cluster rendering task. Finally, we investigate via simulation results the impacts of the job hunger and the resource fragmentation issue of the traditional job scheduling strategies on the system performance, the impacts of the multi-progress and multi-threading cluster rendering on the job completion time and the resource efficiency of the irreversible data dynamic cluster. We further study the extension of the cluster computation capability and the reliability issue of the cloud service.

IP-Bits in Letters 26 December 2018 Self-Correcting Fragmentation in Economics and State Governance (Presentation Slides)

In economics or political life, we often encounter the phenomenon of domination. A dominating position of the company on the market manifests itself via the ability to prevent effective competition between the market players. In politics, the domination of certain political entities has power to limit effective political choice for members of the society. This article will argue that a natural and self-correcting mechanism for dealing with domination is a fragmentation of economic and political resources. A fragmentation of economic and political resources might become a mechanism to regulate the cooperation agreements which are difficult to implement or enforce in practice.

Modeling Service Applications for Optimal Parallel Embedding

Leveraging the traditional virtual network concept, some recent research works have proposed the Virtual Data Center (VDC) as an abstraction to capture both bandwidth and compute/storage resource requirements for an application. However a virtual node in a VDC is limited to a virtual machine (VM), which can only be embedded onto a single physical machine. This condition limits the applicability of the VDC abstraction and the potential of deploying parallel computing. In this paper, we propose a new abstraction based on our Application Centric Network Virtualization (ACNV) approach. Specifically, we model a service application offered by a service provider as a virtual network of service function nodes, which closely matches the service provider's view on the architecture of the application. An infrastructure provider that hosts the application decides how to map the virtual network to the substrate network. Different from the VDC abstraction, each virtual node in our proposed abstraction can be split and mapped onto multiple physical machines, which allows the infrastructure provider to provide auto scaling for the application with variable number of physical machines for exploring the full benefits of parallel computing. We also allow multiple virtual nodes to be mapped and colocated in the same physical machine to minimize resource fragmentation and communication overhead. Extensive simulation results show that the proposed ACNV abstraction outperforms existing VDC-like approaches in achieving optimal resource usage.

A virtual cluster embedding approach by coordinating virtual network and software-defined network

Virtual cluster, as a fundamental service of cloud computing, is an important delivery model of cloud services. Allocating physical resources for a virtual cluster is known as virtual cluster embedding (VCE), which has a significant impact on the performance. VCE includes the components of virtual machine, virtual switch, virtual link, virtual software-defined network (SDN) controller and cluster system controller. VCE needs to be considered comprehensively to adapt the influencing factor weight proportion and to be flexible for different types of embedding. This paper, based on the topology of the cloud data center network, develops a coordinated VCE approach, called CoVCE. The approach combines virtual network with SDN to form virtual SDN and weakens slightly the constraints of VM placement. The network centrality, correlation property and resource fragmentation are optimized with multiple objectives to receive more requests, increase the throughput and decrease network delay and runtime. The CoVCE method integrates not only the logical topology of virtual machines in the cluster but also the relationship with virtual switch, virtual link and control services. This does not only optimize the placement location of virtual components, but also allows the virtual resources to migrate upon virtual cluster requests. The method further improves the utilization of physical resources and reduces resource fragmentation. According to extensive simulation and emulation experiments and comparison with correlative algorithms, CoVCE effectively reduces network delays, offers a higher embedding efficiency, improves user experiences and, to some extent, also improves the revenue/cost ratio and throughput.

Towards improved health service quality in Tanzania: An approach to increase efficiency and effectiveness of routine supportive supervision.

Effective supportive supervision of healthcare services is crucial for improving and maintaining quality of care. However, this process can be challenging in an environment with chronic shortage of qualified human resources, overburdened healthcare providers, multiple roles of district managers, weak supply chains, high donor fragmentation and inefficient allocation of limited financial resources. Operating in this environment, we systematically evaluated an approach developed in Tanzania to strengthen the implementation of routine supportive supervision of primary healthcare providers. The approach included a systematic quality assessment at health facilities using an electronic tool and subsequent result dissemination at council level. Mixed methods were used to compare the new supportive supervision approach with routine supportive supervision. Qualitative data was collected through in-depth interviews in three councils. Observational data and informal communication as well as secondary data complemented the data set. Additionally, an economic costing analysis was carried out in the same councils. Compared to routine supportive supervision, the new approach increased healthcare providers’ knowledge and skills, as well as quality of data collected and acceptance of supportive supervision amongst stakeholders involved. It also ensured better availability of evidence for follow-up actions, including budgeting and planning, and higher stakeholder motivation and ownership of subsequent quality improvement measures. The new approach reduced time and cost spent during supportive supervision. This increased feasibility of supportive supervision and hence the likelihood of its implementation. Thus, the results presented together with previous findings suggested that if used as the standard approach for routine supportive supervision the new approach offers a suitable option to make supportive supervision more efficient and effective and therewith more sustainable. Moreover, the new approach also provides informed guidance to overcome several problems of supportive supervision and healthcare quality assessments in low- and middle income countries.

Fine-grained scheduling in multi-resource clusters

In multi-resource clusters, many schedulers allocate resources based on fixed quantities. However, fixed allocations can easily lead to resource fragmentation and over-commitment problems, which may result in lower resource utilization and performance degradation. This paper proposes a fine-grained method (FGM) to improve the allocation granularity of resource allocation. This method divides tasks into execution stages according to the task requirement estimated using similar tasks at the runtime. Then, task resource requirements are matched with the available server resources by stages to refine two aspects of allocation granularity: allocation duration and allocation quantity. In addition, the FGM may over-allocate resources deliberately to further improve resource utilization and performance. The paper tested the FGM in three environments using both online and offline workloads. The test results show that the FGM can resolve resource fragmentation and over-commitment problems by significantly improving resource utilization and performance with acceptable fairness and scheduling response times.

Strategy-Proof Mechanism for Provisioning and Allocation Virtual Machines in Heterogeneous Clouds

In this paper, we address the problem of heterogeneous physical machines resource management (HPMRM); that is, providing and allocating multiple virtual machine (VM) instances from heterogeneous physical machines to maximize social welfare. Although existing allocation mechanisms allocate VMs to users through the single-mapping mechanism, such allocations cannot guarantee maximum social welfare or efficient utilization of multiple types of resources for cloud providers. Thus, we consider the multi-mapping mechanism, which permits mapping VMs allocated to one user to physical machines for VM provisioning and allocation. This can result in improved social welfare and lead to less resource fragmentation. We formulate the HPMRM problem in an auction-based setting, and design optimal and approximate mechanisms to solve it. In addition, we show that our proposed mechanism is strategy-proof; that is, our proposed mechanism drives the system into an equilibrium where no users have incentives to maximize their own profit by untruthfully reporting their requests. Furthermore, we analyze the approximation ratio of our proposed approximation algorithm. We also perform experiments to investigate the performance of our proposed approximation mechanism compared to the optimal mechanism. Experimental results demonstrate that our proposed approximation mechanism can obtain near optimal solutions and significantly improve allocation efficiency, while generating greater social welfare.

Bird molting and breeding in an area undergoing re-vegetation in the Atlantic Forest of southeastern Brazil

Molting and breeding cycles are high-energetic costs events, and usually synchronized and temporally well defined in tropical avian communities. However, in degraded and/or undergoing restoration areas, environmental stresses, such as forest fragmentation and restriction of food resources, could locally change these patterns. This study aims to analyze molting and breeding in a bird assembly in an undergoing re-vegetation area in the Atlantic Forest of southeastern Brazil. By mist netting, we searched for evidence of molting of fight feathers and presence of brood patch in captured birds. We also recorded three environmental variables (temperature, rainfall, and photoperiod) to test whether they affect those events. We sampled 85 individuals from 36 species. The breeding cycle started in July and peaked in November and December. The molting period was between July and April of the next year and peaked in January and February. Molting and breeding events were related to temperature. The overlap between these events was of 11% (n = 9). Our data suggest that molt and breeding times are in accordance to other studies where the climate is more seasonal, yet more studies are necessary to investigate those biological cycles and their possible alterations due to environmental degradation.

Efficient P‐cycle combination protection strategy based on improved genetic algorithm in elastic optical networks

In order to address the problems of spectrum fragmentation and high redundancy of protection resources in the pre-configured cycle (P-cycle) protection, a genetic P-cycle combination protection strategy (GPCPS) based on improved genetic algorithm (IGA) is proposed for elastic optical networks. In GPCPS, an IGA is devised to optimise the basic cycles and find out an efficient P-cycle protection combination for protecting the whole optical topology. For alleviating the spectrum fragmentation, a P-cycle configuration strategy, named link spectrum partition and P-cycle configuration, is introduced to pre-configure the spectrum resource of P-cycles. Meanwhile, the least traffic priority and first fit algorithm based on spectrum resources usage and path length aware is designed to select working path effectively. Simulation results show that the proposed GPCPS can reduce the bandwidth blocking probability and improve the spectrum resource utilisation.

Interconnection Structures, Management and Routing Challenges in Cloud-Service Data Center Networks: A Survey

Today’s data center networks employ expensive networking equipments in associated structures that were not designed to meet the increasing requirements of the current large-scale data center services. Limitations that vary between reliability, resource utilization, and high costs are challenging. The era of cloud computing represents a promise to enable large-scale data centers. Computing platforms of such cloud service data centers consist of large number of commodity low-price servers that, with a theme of virtualization on top, can meet the performance of the expensive high-level servers at only a fraction of the price. Recently, the research in data center networks started to evolve rapidly. This opened the path for addressing many of its design and management challenges, these like scalability, reliability, bandwidth capacities, virtual machines’ migration, and cost. Bandwidth resource fragmentation limits the network agility, and leads to low utilization rates, not only for the bandwidth resources, but also for the servers that run the applications. With Traffic Engineering methods, managers of such networks can adapt for rapid changes in the network traffic among their servers, this can help to provide better resource utilization and lower costs. The market is going through exciting changes, and the need to run demanding-scale services drives the work toward cloud networks. These networks that are enabled by the notation of autonomic management, and the availability of commodity low-price network equipments. This work provides the readers with a survey that presents the management challenges, design and operational constraints of the cloud-service data center networks

A Provident Resource Defragmentation Framework for Mobile Cloud Computing

To facilitate mobile cloud computing, a cloud service provider must dynamically create and terminate a large number of virtual machines (VMs), causing fragmented resources that cannot be further utilized. To solve this problem proactively, most of the existing studies have been based on server consolidation, with the main objective of minimizing the number of active servers. Although this approach can minimize resource fragmentation at a particular time, it may be over aggressive at the price of too frequent VM migration and low system stability. To address this issue, we propose a novel provident resource defragmentation framework that is revenue-oriented with the goal to reduce unnecessary VM migration. Within the proposed framework, we formulate an optimization problem for resource defragmentation at a particular time epoch, with the consideration of the future impact of any VM migration. We then develop an efficient heuristic algorithm that can obtain near-optimal results. Extensive numerical results confirm that our framework can provide the highest profit and can significantly reduce the VM migration cost in practical scenarios.

Collaborative Dynamic Virtual Network Embedding Algorithm Based on Resource Importance Measures

Many existing virtual network embedding (VNE) algorithms adopt the greedy embedding strategies, which embed the in-progress virtual network requests (VNRs) onto the substrate nodes and links with more residual resources. However, the previous VNRs may overconsume the critical substrate resources and lead to the resource fragmentation problem in the substrate networks, thus reducing the acceptance ratio of the latter VNRs. To address this problem, we propose a novel collaborative VNE algorithm based on resource importance measures, RIM-ViNE, in which the global node importance and link importance in the initial substrate network are measured using multiple topological attributes and are used to set the embedding cost of VNRs. Then, the VNE problem is defined as a linear programming model and is solved by minimizing the total embedding cost, which could bring about the coordinative embedding between different VNRs, and thus preventing the critical resources from being over occupied by the previous VNRs with small requirements and improving the resource fragmentation problem. Moreover, we propose a dynamic reconfiguration mechanism based on critical nodes protection (CNP-Re) to improve the resource fragmentation problem further. Extensive experiments are conducted under two network scenarios, and the results show that the proposed algorithms outperform the VNE algorithms that only consider the coordination between node mapping and link mapping or only measure the node importance in the residual substrate network, and the average VNR acceptance ratio, average revenue, and average resource utilization ratio are effectively improved.

A Generalized VNF Sharing Approach for Service Scheduling

Network function virtualization enables a flexible service provision by introducing the concept of virtual network function (VNF). Despite the importance of VNF scheduling, it is largely unexplored. In this letter, we propose a new model for VNF scheduling based on the min-plus algebra theory. The main idea of this model is to share the deployed VNF instances among different services, thus to improve the resource utilization and reduce the resource fragmentation generated by deploying many VNF instances. In addition, a weight-based VNF sharing scheduling approach is proposed in this model to achieve a fair scheduling among the arriving service requests. The experimental results show that the proposed model and approach are effective and efficient.

Constraints and capacities for novel livelihood adaptation: lessons from agricultural adoption in an African dryland pastoralist system

Climate change, land use change, and sociopolitical and institutional transitions in African drylands have resulted in the loss, fragmentation, and degradation of environmental resources that pastoralists rely upon to sustain their livestock-based livelihoods. Diversification into irrigated agriculture is a potential strategy to increase food security. However, successful livelihood adaptation depends on access to key forms of natural, human, social, financial, and physical capitals, which may be lacking or unequally distributed in pastoralist communities. In two dryland pastoralist communities in Kenya, an international relief organization introduced irrigated farming in 2010. Nearly 200 individuals began farming, but within 4 years, all had quit. This study investigated the role of household access to different forms of capital in decisions to adopt agriculture, and how adaptive capacities, environmental conditions, and farming as a new livelihood practice interacted to shape household vulnerability. We found that decisions to farm were largely decoupled from access to assets that would afford greater adaptive capacity. The analysis also identified critical constraints that households encountered once they adopted farming. This study highlights the importance of more nuanced understandings of local capacities in the dynamics of adapting to environmental change. We recommend that prior to introducing interventions that promote new livelihood activities, development organizations should assess local capacities and pursue targeted strategies to increase household-level access to the forms of assets that will facilitate successful adaptation.

Provider Perspectives on Intimate Partner Violence in Bluefields, Nicaragua

ABSTRACTBackground: Intimate partner violence represents a significant public health problem and a substantial human rights' issue for women and girls throughout the world. Design and Purpose: The purpose of this study was to answer these research questions: What are the current practices for addressing gender violence in the RACS? What do professionals consider to be the current strengths and gaps in policies related to gender violence in this region? By employing a qualitative descriptive approach (Sandelowski, 2000), researchers traveled from the US to Bluefields, Nicaragua, in 2012. The multidisciplinary team of two US nurses, a prosecutor, and a victim-witness advocate interviewed 18 key informants, police officers, advocates, and nurses, and observed court processes. Methods: Interviews were transcribed verbatim and analyzed in the language the interview was conducted in. Researchers coded data independently and identified emergent themes. Findings: Informants described the complexity of the nature and dynamics of gender violence, strongly informed by Nicaragua's fairly progressive laws. The participants described holistic, fully integrated services as the intended functioning of the system. These services were often thwarted by gaps—fragmentation and lack of resources—and were additionally hampered by substantial individual and structural economical obstacles.

Product quality asymmetry and food safety: Investigating the “one farm household, two production systems” of fruit and vegetable farmers in China

The Chinese government faces a great challenge to safeguard the safety of the food supply chain due to the significant fragmentation of producers and weak institutional resources to monitor and enforce safety standards. Food safety in China has received escalating attention since the 2008 milk scandal. Mainly due to safety concerns, Chinese farmers adopt two separate production systems for the market and self-consumption, and thus, the so-called “One Farm Household, Two Production Systems” (OFH-TPS) has gained popularity in the recent years.This study provides a theoretical framework to examine how product inspections and certifications affect the OFH-TPS adoption by mitigating information asymmetry of product quality. It also shows that both food safety concerns and cost savings are important factors for famers' decision on the adoption. The empirical analysis finds that product inspections conducted by firms, agricultural cooperatives, and farmers curb the OFH-TPS adoption, whereas government inspections have no statistically significant effect. Farmers who grow “Green Food” are less likely but organic farmers are more likely to adopt the OFH-TPS. We also find that pesticide use training curbs the OFH-TPS adoption, and the perceived adverse effects of pesticide use to the environment and human health have no statistical effect. Farmers who use highly toxic or banned pesticides and/or who perceive food safety at the local markets is poor are more likely to adopt the OFH-TPS. This study provides rich policy implications. To safeguard food safety, it is critical to engage the private sector for co-regulation, improve the efficacy of government inspections, and promote education on pesticide use to both farmers as well as retailers of pesticides.

Improving performance by network-aware virtual machine clustering and consolidation

Modern data center consists of thousands of servers, racks and switches. Complicated structure means it requires well-designed algorithms to utilize resources of data centers efficiently. Current virtual machine scheduling algorithms mainly focus on the initial allocation of virtual machines based on the CPU, memory and network bandwidth requirements. However, when tasks finished or lease expired, related virtual machines would be deleted from the system which would generate resource fragments. Such fragments lead to unbalanced resource utilization and decline of communication performance. This paper investigates the network influence on typical applications in data centers and proposed a self-adaptive network-aware virtual machine clustering and consolidation algorithm to maintain an optimal system-wide status. Our consolidation algorithm periodically checks whether consolidation is necessary and then clusters and consolidates virtual machines to lower communication cost with an online heuristic. We used two benchmarks in a real environment to examine network influence on different tasks. To evaluate the advantages of the proposed algorithm, we also built a cloud computing testbed. Real workload trace-driven simulations and testbed-based experiments showed that, our algorithm greatly shortened the average finish time of map-reduce tasks and reduced time delay of web applications. Simulation results showed that our algorithm considerably reduced the amount of high-delay jobs, lowered the average traffic passed through aggregate switches and improved the communication ability among virtual machines.

A framework for simulating large scale cloud infrastructures

Cloud infrastructures are continuously growing in size, since more cloud nodes are added to already existing hyper-scale infrastructures. These hyper-scale infrastructures are also becoming heterogeneous as different types of accelerators are added in order to increase performance per watt for certain types of applications and allow for various HPC workloads to migrate to Cloud environments. The introduction of diverse workloads that migrate in the Cloud along with increasing volume of incoming tasks results in phenomena of network congestion, underutilization and resource fragmentation. Simulators are used to analyze, study and possibly improve Cloud environments. However, existing Cloud simulation tools lack the ability to handle heterogeneous resources and tasks that span across multiple Cloud nodes. Moreover, they are mostly sequential and cannot scale to large numbers of Cloud nodes. Furthermore, they do not support over-commitment, which is a common practice in real-world Cloud environments. A framework for simulating large numbers of heterogeneous cloud nodes organized in Cells and executing large numbers of HPC tasks is proposed. The framework is inherently parallel and designed for hybrid distributed memory parallel systems, supporting CPU, memory and network over-commitment. The simulation framework is based on a time advancing loop, allowing dynamic change of the granularity of the simulator and minimizing memory requirements, since data related to the current time-step is stored. Moreover, a latency model for the currency of data in the Gateway Service and Broker is also supported. Implementation details along with discussions concerning the extensibility of the framework are given. Numerical results for simulating large number of heterogeneous resources and incoming tasks are also presented.