Service IaaS Research Articles

Monitoring of IaaS and scientific applications on the Cloud using the Elasticsearch ecosystem

The private Cloud at the Torino INFN computing centre offers IaaS services to different scientific computing applications. The infrastructure is managed with the OpenNebula cloud controller. The main stakeholders of the facility are a grid Tier-2 site for the ALICE collaboration at LHC, an interactive analysis facility for the same experiment and a grid Tier-2 site for the BES-III collaboration, plus an increasing number of other small tenants. Besides keeping track of the usage, the automation of dynamic allocation of resources to tenants requires detailed monitoring and accounting of the resource usage. As a first investigation towards this, we set up a monitoring system to inspect the site activities both in terms of IaaS and applications running on the hosted virtual instances. For this purpose we used the Elasticsearch, Logstash and Kibana stack. In the current implementation, the heterogeneous accounting information is fed to different MySQL databases and sent to Elasticsearch via a custom Logstash plugin. For the IaaS metering, we developed sensors for the OpenNebula API. The IaaS level information gathered through the API is sent to the MySQL database through an ad-hoc developed RESTful web service, which is also used for other accounting purposes. Concerning the application level, we used the Root plugin TProofMonSenderSQL to collect accounting data from the interactive analysis facility. The BES-III virtual instances used to be monitored with Zabbix, as a proof of concept we also retrieve the information contained in the Zabbix database. Each of these three cases is indexed separately in Elasticsearch. We are now starting to consider dismissing the intermediate level provided by the SQL database and evaluating a NoSQL option as a unique central database for all the monitoring information. We setup a set of Kibana dashboards with pre-defined queries in order to monitor the relevant information in each case. In this way we have achieved a uniform monitoring interface for both the IaaS and the scientific applications, mostly leveraging off-the-shelf tools.

Automatic verification technology of software patches for user virtual environments on IaaS cloud

We propose here a technique for automatic verification of software patches for user virtual environments on Infrastructure as a Service (IaaS) Cloud to reduce the cost of verifying patches. IaaS services have been spreading rapidly, and many users can customize virtual machines on IaaS Cloud like their own private servers. However, users must install and verify software patches of the OS or middleware installed on virtual machines by themselves. This task increases the user’s operation costs. Our proposed method replicates user virtual environments, extracts verification test cases for user virtual environments from a test case database (DB), distributes patches to virtual machines in the replicated environments, and executes the test cases automatically on the replicated environments. To reduce test cases creation efforts, we propose an idea of two-tier abstraction which groups software to software groups and function groups and selects test cases belonging to each group. We applied the proposed method on OpenStack using Jenkins and confirmed its feasibility. We evaluated the effectiveness of test case creation efforts and the automatic verification performance of environment replications, test cases extractions, and test case executions.

SLA-BASED RESOURCE ALLOCATION WITHIN CLOUD NETWORKING ENVIRONMENT

Today, cloud networking which is the ability to connect users with their cloud services and to interconnect these services in an inter-cloud approach, is one of the recent research areas within the cloud computing research communities. The main drawback of cloud networking consists in the lack of Quality of Service (QoS) guarantee and management in conformance with a corresponding Service Level Agreement. In this paper, a framework is proposed for resource allocation according to an end-to-end service level agreement established between a Cloud Service User and several Cloud Service Providers within a cloud networking environment. The paper focuses on QoS parameters for NaaS and IaaS services. In addition, several algorithms are proposed for the best Cloud Service Providers selection to allocate VMs and network resources within a Broker and a Federation based inter-cloud scenarios. The objective is to minimize the total cost while satisfying NaaS and IaaS QoS constraints. Moreover, the proposed cloud networking architecture is used to provide videoconferencing and intensive computing applications with QoS guarantee. The corresponding scenarios are simulated using an extended CloudSim toolkit. The results reveal a high scalability and good performances where the Broker architecture is the most interesting while ensuring QoS requirements.

Cloud Federations in the Sky: Formation Game and Mechanism

The amount of computing resources required by current and future data-intensive applications is expected to increase dramatically, creating high demands for cloud resources. The cloud providers' available resources may not be sufficient enough to cope with such demands. Therefore, the cloud providers need to reshape their business structures and seek to improve their dynamic resource scaling capabilities. Federated clouds offer a practical platform for addressing this service management issue. We introduce a cloud federation formation game that considers the cooperation of the cloud providers in offering cloud IaaS services. Based on the proposed federation formation game, we design a cloud federation formation mechanism that enables the cloud providers to dynamically form a cloud federation maximizing their profit. In addition, the proposed mechanism produces a stable cloud federation structure, that is, the participating cloud providers in the federation do not have incentives to break away from the federation. We analyze the performance of the proposed mechanism by performing extensive experiments. The results of the experiments show that the cloud federation obtained by our proposed mechanism is stable, yielding high profit for the participating cloud providers.

Development of resource management server for production IaaS services based on OpenStack

Development of resource management server for production IaaS services based on OpenStack

Agents-based deployment of heterogeneous IaaS clouds

Cloud technology has provided much more effective and flexible management of computing infrastructure both to service vendors and customers. However, the provisioning of computing resources and deployment of applications using heterogeneous infrastructure as a service IaaS clouds is not straight forward, because offers are not easily comparable, providers ask for using different technologies and a standard is still lacking. Moreover cloud providers often hinder the adoption of a common standard in order to lock customers to their solutions. To address these issues we have designed and developed, within the research activities of the mOSAIC project, an agents-based solution that provides a uniform interface to IaaS provisioning and management. The Cloud Agency uses a set of vendor agents implementing adapters to different commercial cloud providers and open source cloud technologies. Starting from the computational constraints and from the application requirements, the vendor agents allow the users to negotiate the best IaaS solution that is available in the cloud market and to build and deploy his/her cloud infrastructure.

On the Conceptualization of Performance Evaluation of IaaS Services

Cloud Computing has been increasingly accepted as a promising computing paradigm in industry, with one of the most common delivery models being Infrastructure as a Service (IaaS). An increasing number of providers have started to supply public IaaS services with different terminologies, definitions, and goals. As such, understanding the full scope of performance evaluation of candidate services would be crucial and beneficial for both service customers (e.g., cost-benefit analysis) and providers (e.g., direction of improvement). Given the numerous and diverse IaaS service features to be evaluated, a natural strategy is to implement different types of evaluation experiments separately. Unfortunately, it could be hard to fairly distinguish between different experimental types due to different environments and techniques that may be adopted by different evaluators. To overcome such obstacles, we have first established a novel taxonomy to help profile and clarify the nature of IaaS services performance evaluation and then built a three-layer conceptual model to generalize the existing performance evaluation practices. Using relevant elements/classifiers in the taxonomy and conceptual model, evaluators can construct natural language-style descriptions and experimental design blueprints to outline the evaluation scope and also to guide new evaluation implementations. In essence, the generated descriptions and blueprints abstractly define and characterize the actual evaluation work. This enables relatively fair and rational comparisons between different performance evaluations according to their abstract characteristics.

OCSO: Off-the-cloud service optimization for green efficient service resource utilization

Many efforts have been made in optimizing cloud service resource management for efficient service provision and delivery, yet little research addresses how to consume the provisioned service resources efficiently. Meanwhile, typical existing resource scaling management approaches often rest on single monitor category statistics and are driven by certain threshold algorithms, they usually fail to function effectively in case of dealing with complicated and unpredictable workload patterns. Fundamentally, this is due to the inflexibility of using static monitor, threshold and scaling parameters. This paper presents Off-the-Cloud Service Optimization (OCSO), a novel user-side optimization solution which specifically deals with service resource consumption efficiency from the service consumer perspective. OCSO rests on an intelligent resource scaling algorithm which relies on multiple service monitor metrics plus dynamic threshold and scaling parameters. It can achieve proactive and continuous service optimizations for both real-world IaaS and PaaS services, through OCSO cloud service API. From the two series of experiments conducted over Amazon EC2 and ElasticBeanstalk using OCSO prototype, it is demonstrated that the proposed approach can make significant improvement over Amazon native automated service provision and scaling options, regardless of scaling up/down or in/out.

Trusted Virtual Machine Allocation in Cloud Computing IaaS Service

Cloud computing is a new era in computing paradigm. It helps Information Technology (IT) companies to cut the cost by outsourcing data and computation on-demand. Cloud computing provides different kind of services which includes Hardware as a Service, Software as a Service (SaaS), Infrastructure as a Service (IaaS) etc. Despite these potential benefits, many IT companies are reluctant to do cloud business due to outstanding trust issues. Cloud consumer and provider are the most interested parties to maximize their benefits. In IaaS, the cloud provider operates the whole computing platform as a resource for the customer, which is accessed by customer as a Virtual Machine (VM) via the internet. The cloud provider must predict the best machine among the available machines to launch VM. This strategic prediction would avoid exodus of computation in middle due to machine heavy load or any failure which severely affect the benefits of both consumer and provider. Since VM allocation for IaaS request is a challenging task, in this study novel architecture is proposed for IaaS cloud computing environment in which VM allocation is done through genetically weight optimized neural network. In this scenario the host load of each machine is taken as its resource information. The neural network predicts the host load of each machine in near future based on the recent past host load. It would help the VM allocator to choose the right machine. Analysis is done on the performance of genetically weight optimized Back Propagation Neural Network (BPNN), Elman Neural Network (ELNN) and Jordan Neural Network (JNN) for prediction accuracy.

An ontology update mechanism in IaaS service discovery system

Purpose – Hybrid cloud computing is considered a viable and cost-effective approach to satisfy the inability of a private cloud to meet the user requirements. The information status of the selected public cloud service may change at runtime which should be reflected at the broker server database. This research illustrates a mechanism to assist the IaaS discovery system to assess the status change of the selected public cloud service and to update the broker server database. The paper aims to discuss these issues. Design/methodology/approach – A prototype was developed with the broker server as the main component in the service discovery system containing the status information of the selected public cloud service. The merge-ontology and patch-update methods were proposed, the processing cost details of each were measured, and the methods were evaluated. Findings – Experimental results showed that in the merge update, the merging process incurred much longer time than its required communication, contributing to long overall time. Relatively, the patch update method incurred much less time than its counterpart. Research limitations/implications – The proposed mechanism is experimental with some ideal assumptions, and so further work in real conditions is needed for its improvement. Originality/value – This research is believed to be the first proposal to investigate ontology merge/patch methods to support ontology update in the broker server database of a hybrid cloud and will serve as a reference to researchers in the field.

A New Approach to Manage and Utilize Cloud Computing Underused Resources

In this paper a new approach for driving a better cloud computing IaaS Services is presented. This approach focuses on extending the available cloud computing platform infrastructure by harvesting underused generic computing resources that are widely available within public domains such as universities and organizations. Most of the current cloud computing platforms are mainly based on using a dedicating infrastructure to achieve the requested services. The proposed approach aims to improve the computing power for the cloud computing platform without charging any extra cost since generic machines are owned by the enterprise. A new Resources management mechanism is introduced to manage the combination of the dedicated and generic machines. In order to implement and achieve the goals of the proposed approach several challenges should be conquered. These challenges are coming from the inherently stochastic characteristics of the harvested unused computing such as reliability and hardware compatibility. In the implementation, Openstack cloud computing platform is used and extended in such a way that guarantees QoS and an opportunistic use of the idle or underused generic or public computing resources.

Cost models - pillars for efficient cloud computing: position paper

Cost models are fundamental building blocks for cloud computing. Cloud providers offer a very wide portfolio of services, while Cloud clients access them against some financial arrangement. There is a fundamental trade-off between what Cloud provider can offer in terms of resources software or hardware, services e.g., storage, e-mail cost and what Cloud clients are willing to pay. The Cloud cost models today are distinct based on service models: Infrastructure as a Service IaaS, Platform as a Service PaaS and Software as a Service SaaS. The need to forecast the cost over a period of time imposes building of cost models, which have to be accurate and error free. In this paper, we make a survey and analyse existing cost models in cloud computing and discuss open issues related to the subject.

Game Theory Based Iaas Services Composition in Cloud Computing Environment

Game Theory Based Iaas Services Composition in Cloud Computing Environment

Applying security policies and service level agreement to IaaS service model to enhance security and transition

Over a decade ago, cloud computing became an important topic for small, medium and large businesses alike. The new concept promises scalability, security, cost reduction, portability and availability. While addressing the cloud concepts over the past several years, there have been intensive discussions about the importance of the different cloud computing service model. Moreover, there were lots of discussions about the risks in migrating to cloud computing. Therefore, this paper reviews the concept of cloud computing, security policies and concentrates on Infrastructure as a Service (IaaS) model. Also, the paper examines the risks encountered by implementing the Infrastructure as a Service (IaaS) model in organizations. Furthermore, the paper’s aim is to discuss the role of security policies, service level agreement (SLA) and compliance for enhancing the security of the IaaS service model by presenting several applicable policies.

Toward role-based provisioning and access control for infrastructure as a service (IaaS)

Abstract Cloud computing has drawn much attention in recent years. One of its service models, called infrastructure as a service (IaaS), provides users with infrastructure services such as computation and data storage, heavily dependent upon virtualization techniques. Most of the current IaaS providers take the user-resource direct mapping approach for their business, where individual users are the only type of service consumer who can request and use virtualized resources as long as they pay for the usage. Therefore, in this approach, the users and virtual resources are centrally managed at the IaaS providers. However, this also results in the lack of support for scalable authorization management of users and resources, organization-level policy support, and flexible pricing for business users. Considering the increasing popularity and growing user base of cloud computing, there is a strong need for a more flexible IaaS model with a finer grained access control mechanism than the aforementioned all-or-nothing approach. In this paper we propose a domain-based, decentralized framework for provisioning and managing users and virtualized resources in IaaS. Specifically, an additional layer called domain is introduced to the user-resource direct mapping scheme, whereby de-centralization of user and resource management is facilitated. Our framework also allows the IaaS service provider to delegate its administrative routines to domains so that each domain is able to manage its users and virtualized resources allocated by the IaaS provider. Our domain-based approach offers benefits such as scalable user/resource management, domain-based security and governance policy support, and flexible pricing.