Economics of Grids, Clouds, Systems, and Services

Typical pricing models for IaaS cloud providers are slotted, using hour and month as time units for metering and charging resource usage. Such models lead to financial loss as applications may release resources much earlier than the end of the last allocated time slot, leaving the cost paid for the rest of the time unit wasted. This problem can be minimized for multi-tenant environments by managing resources as pools. This scenario is particularly interesting for universities and companies with various departments and SaaS providers with multiple clients. In this paper we introduce a tool that creates and manages resource pools for multi-tenant environments. Its benefit is the reduction of resource waste by reusing already allocated resources available in the pool. We discuss the architecture of this tool and demonstrate its effectiveness, using a seven-month workload trace obtained from a real multi-tenant SaaS financial risk analysis application. From our experiments, such tool reduced resource costs per day by 13 % on average in comparison to direct allocation of cloud provider resources.

In this paper we handle the problem of scheduling tasks in hybrid clouds for small companies which can spend only a fixed budget in order to handle specific situations where the demand is high and cannot be predicted. We describe a model with important characteristics for the resource utilization and we design an algorithm for scheduling tasks which are sent continuously for execution, optimizing the schedule for tasks with high priority and short deadline. We propose an architecture that meets the challenges encountered by small business in their systems for tasks scheduling. We describe the main components, Configuration Agent and Task Scheduler, and we analyze different test scenarios, proving the efficiency of the proposed strategy.

In recent years, cloud computing providers have been working to provide highly available and scalable cloud services to keep themselves alive in the competitive market of various cloud services. The difficulty is that to provide such high quality services, they need to enlarge data centers (DCs), and consequently, to increase operating costs. Hence, leveraging cost-aware solutions to manage resources is necessary for cloud providers to decrease the total energy consumption, while keeping their customers satisfied with high quality services. In this paper, we consider the cost-aware virtual machine (VM) placement across geographically distributed DCs as a multi-criteria decision making problem and propose a novel approach to solve it by utilizing Bayesian Networks and two algorithms for VM allocation and consolidation. The novelty of our work lays in building the Bayesian Network according to the extracted expert knowledge and the probabilistic dependencies among parameters to make decisions regarding cost-aware VM placement across distributed DCs, which can face power outages. Moreover, to evaluate the proposed approach we design a novel simulation framework that provides the required features for simulating distributed DCs. The performance evaluation results reveal that using the proposed approach can reduce operating costs by up to 45 % in comparison with First-Fit-Decreasing heuristic method as a baseline algorithm.

Cloud providers now offer resources as combinations of CPU frequencies and prices, with faster resources (which operate at higher frequencies) charged at a higher monetary cost. With the emergence of this new pricing scheme, the problem of choosing cost-efficient configurations is becoming even more challenging for users. The frequencies required to achieve cost-efficient configurations may vary in different scenarios, depending on both the provider’s pricing model and the application characteristics. In this paper, two cost-aware algorithms that select low-cost CPU frequencies for each resource to complete a scientific workflow application within a deadline and at a minimum cost are presented. The proposed approaches are evaluated and compared through simulation using different pricing models that charge resource provisioning also based on the CPU frequency.

The economic cost is a decisive factor that influences the migration of an application to a cloud infrastructure. Once the migration has been decided, the cost of cloud resources that will be hired to run the application must be minimized considering the application and user constraints. In this paper, we propose a method to determine the cheapest combination of computing instances to execute bag-of tasks applications in the Amazon Elastic Compute Cloud (EC2) infrastructure. The method considers the heterogeneity of the resources (types of computing instances, regions and availability zones, or purchasing and payment options) as well as the deadline and the input workload provided by the user. The paper also shows how the proposed method improves provisioning decisions previously adopted by the authors to execute a linked-data application.

A particular problem of cloud environments is the assurance of a certain level of service quality. The problem is escalated in the case of building support platforms for using multiple clouds. Various partial solutions to ensure a certain quality level of the cloud services have been investigated in the last half decade. This paper analyzes the existing approaches to define, model, evaluate, estimate, measure of optimize the quality of services offered to cloud-based applications. A particular approach is detailed, the one that uses model-driven engineering techniques. Moreover, the special case of designing data-intensive applications, the appropriate quality of service attributes are identified.

Cloud computing is essentially changing the way services are built, provided and consumed. As a paradigm building on a set of combined technologies, it enables service provision through the commoditization of IT assets and on-demand usage patterns. In the emerging era of the Future Internet, clouds aim at facilitating applications that move away from the monolithic approach into an Internet-scale one, thus exploiting information, individual offerings and infrastructures as composite services. In this paper we present an approach for selecting the services (that comprise the composite ones) in order to meet the end-to-end Quality of Service (QoS) requirements. The approach is enhanced with a relevance feedback mechanism that provides additional information with respect to the importance of the content and the service. The latter is performed in an automated way, allowing for user preferences to be considered during the service selection process. We also demonstrate the operation of the implemented approach and evaluate its effectiveness using a real-world scenario, based on a computer vision application.

When trying to discover, assess, and select cloud services, companies face many challenges, such as fast-moving markets, vast numbers of offerings, and highly ambiguous selection criteria. This publication presents the Open Service Compendium (OSC), an information system which supports businesses in their discovery, assessment and cloud service selection by offering a simple dynamic service description language, business-pertinent vocabularies, as well as matchmaking functionality. It contributes to the state of the art by offering a more practical, mature, simple, and usable approach than related works.

In this paper a methodology for optimizing Data Centres (DCs) operation allowing them to provide various types of Ancillary Services on-demand is proposed. Energy flexibility models have been defined for hardware devices inside DCs aiming at optimizing energy demand profile by means of load time shifting, alternative usage of non-electrical cooling devices (e.g. thermal storage) or charging/discharging the electrical storage devices. As result DCs are able to shape their energy demand to provide additional load following reserve for large un-forecasted wind ramps, shed or shift energy demand over time to avoid an coincidental peak load and feed back in the grid the energy produced by turning on their backup fossil fuelled generators to maintain (local) reactive power balance under normal conditions. Experiments via numerical simulations based on real world traces of DC operation highlight the methodology potential for optimizing DC energy consumption to provide Ancillary Services.

Requirements of Energy Management Services include short and long term processing of data in a massively interconnected scenario. The complexity and variety of short term applications needs methodologies that allow designers to reason about the models taking into account functional and non-functional requirements. In this paper we present a component based specification language for building trustworthy continuous dataflow applications. Component behaviour is defined by Petri Nets in order to translate to the methodology all the advantages derived from a mathematically based executable model to support analysis, verification, simulation and performance evaluation. The paper illustrates how to model and reason with specifications of advanced data flow abstractions such as smart grids.

As the use of smartphones and its applications continue their rapid growth, prolonging the smartphone battery lifetime has become one of the main concerns for smartphone users if re-charging is not possible. In this paper, we show that, by taking into account the user preferences, the energy consumption of smartphones can be adjusted to maximize the user utility. The user preferences are reflected through the type of application uses, the perceived costs of energy allocation for the different types of applications, and the perceived value of energy remaining in the battery of the smartphone. In particular, we optimize the energy consumption of smartphones through the use of a utility-based energy consumption optimization model, which we developed. We demonstrate the workings of our model by applying it to a simple scenario, in which we vary the perceived value of energy remaining in the smartphone battery and the user’s perceived costs for energy consumed by the two types of application uses: cloud-based application uses and on-device application uses. Our results show that, by letting users express their preferences, users can allocate the remaining smartphone energy such that it maximizes their utilities.

This paper presents a Particle Swarm Optimization-based method for optimizing the energy consumption in data centers. A particle position is mapped on a data center configuration (i.e. allocation of virtual machines on the data center’s servers) which is evaluated using a fitness function that considers the energy consumed by the servers’ hardware resources and by the data center’s cooling system as evaluation criteria. The Particle Swarm Optimization-based method is triggered each time a workload arrives to be accommodated on the data center’s servers. The proposed method has been integrated in the CloudSim framework and has been evaluated on randomly generated logs.

Energy consumption in Cloud computing is a significant issue in regards to aspects such as the cost of energy, cooling in the data center and the environmental impact of cloud data centers. Smart grids offers the prospect of dynamic costs for a data center’s energy usage. These dynamic costs can be passed on to Cloud users providing incentives for users to moderate their load while also ensuring the Cloud providers are insulated from fluctuations in the cost of energy. The first step towards this is an architecture that focuses on energy monitoring and usage prediction. We provide such an architecture at both the PaaS and IaaS layers, resulting in energy metrics for applications, VMs and physical hosts, which is key to enabling active demand in cloud data centers. This architecture is demonstrated through our initial results utilising a generic use case, providing energy consumption information at the PaaS and IaaS layers. Such monitoring and prediction provides the groundwork for providers passing on energy consumption costs to end users. It is envisaged that the resulting varying price associated with energy consumption can help motivate the formation of methods and tools to support software developers aiming to optimise energy efficiency and minimise the carbon footprint of Cloud applications.

The Diameter of Things (DoT) protocol is intended to provide a near real-time metering framework for IoT applications in resource-constraint gateways. Respecting resource capacity constraints on edge devices establishes a firm requirement for a lightweight protocol in support of fine-grained telemetry of IoT deployment units. Such metering capability is needed when lack of resources among competing applications dictates our schedule and credit allocation. In response to these findings, the authors offer the DoT protocol that can be incorporated to implement real-time metering of IoT services for prepaid subscribers as well as Pay-per-use economic models. The DoT employs mechanisms to handle the IoT composite application resource usage units consumed/charged against a single user balance. Such charging methods come in two models of time-based and event-based patterns. The former is used for scenarios where the charged units are continuously consumed while the latter is typically used when units are implicit invocation events. The DoT-enabled platform performs a chained metering transaction on a graph of dependent IoT microservices, collects the emitted usage data, then generates billable artifacts from the chain of metering tokens. Finally it permits micropayments to take place in parallel.

Web applications have become crucial components of current service-oriented business applications. Therefore, it is very important for the company’s reputation that the performance of a web application has been tested thoroughly before deployment. We present a tool-supported performance exploration approach to investigate how potential user behavioral patterns affect the performance of the system under test. This work builds on our previous work in which we generate load from workload models describing the expected behavior of the users. We mutate a given workload model (specified using Probabilistic Timed Automata) in order to generate different mutants. Each mutant is used for load generation using the MBPeT tool and the resource utilization of the system under test is monitored. At the end of an experiment, we analyze the mutants in two ways: cluster the mutants based on the resource utilization of the system under test and identify those mutants that satisfy the criteria of given objective functions.

About 80 % of the financial market investors fail, the main reason for this being their poor investment decisions. Without advanced financial analysis tools and the knowledge to interpret the analysis, the investors can easily make irrational investment decisions. Moreover, investors are challenged by the dynamism of the market and a relatively large number of indicators that must be computed. In this paper we propose E-Fast, an innovative approach for on-line technical analysis for helping small investors to obtain a greater efficiency on the market by increasing their knowledge. The E-Fast technical analysis platform prototype relies on High Performance Computing (HPC), allowing to rapidly develop and extensively validate the most sophisticated finance analysis algorithms. In this work, we aim at demonstrating that the E-Fast implementation, based on the CloudPower HPC infrastructure, is able to provide small investors a realistic, low-cost and secure service that would otherwise be available only to the large financial institutions. We describe the architecture of our system and provide design insights. We present the results obtained with a real service implementation based on the Exponential Moving Average computational method, using CloudPower and Grid5000 for the computations’ acceleration. We also elaborate a set of interesting challenges emerging from this work, as next steps towards high performance technical analysis for small investors.

Community network clouds provide for applications of local interest deployed within community networks through collaborative efforts to provision cloud infrastructures. They complement the traditional large-scale public cloud providers similar to the model of decentralised edge clouds by bringing both content and computation closer to the users at the edges of the network. Services and applications within community network clouds require connectivity to the Internet and to the resources external to the community network, and here the current best-effort model of volunteers contributing gateway access in the community networks falls short. We model the problem of reserving the bandwidth at such gateways for guaranteeing quality-of-service for the cloud applications, and evaluate different pricing mechanisms for their suitability in ensuring maximal social welfare and eliciting truthful requests from the users. We find second-price auction based mechanisms, including Vickrey and generalised second price auctions, suitable for the bandwidth allocation problem at the gateways in the community networks.

The Internet and cloud services are key enablers for participation in society. The need for Internet access in areas underserved by commercial telecom operators has often been a motivation to develop community networks. Many examples around the world show successful cooperative developments of open, participatory local networking infrastructures. Such collaborative models have not yet been applied to local cloud computing resources and services. In this paper, we elaborate on the sustainability model of the http://​guifi.​net community network as a basis for cloud-based infrastructures and services in communities. We first look at the elements of http://​guifi.​net, which support the sustainability and growth of the networking infrastructure. We then discuss their application to cloud-based services within the network and come up with a framework of tools and components for community cloud resources and services. Finally, we assess the current status of the experimental community cloud in http://​guifi.​net, where some of the proposed tools are already operational.

A major part of the commercial Internet is moving towards a cloud paradigm. This phenomenon has a drastic impact on the organizational structures of enterprises and introduces new challenges that must be properly addressed to avoid major setbacks. One such challenge is that of cloud provider viability, that is, the reasonable certainty that the Cloud Service Provider (CSP) will not go out of business, either by filing for bankruptcy or by simply shutting down operations, thus leaving its customers stranded without an infrastructure and, depending on the type of cloud service used, even without their applications or data. This article attempts to address the issue of cloud provider viability, proposing some ways of mitigating the problem both from a technical and from a legal perspective.

In recent studies, Information and Communication Technologies have been key drivers of innovation and economic growth throughout the world. Because the Information and Communication Technology products and services require intensive knowledge, leading countries invested in their innovation systems to operate more effectively and efficiently. Studies on innovation have investigated the knowledge base of countries and their respective relationships with their national institutions, and subsequent economic growth to identify factors which have led to success. However, the approaches of previous studies omit the constituents of the knowledge base while focusing on quantitative aspects such as size. In this article, I propose a novel approach to exploring the knowledge base at a global level by undertaking a network analysis of patents. In this framework, the global knowledge network is defined as a set of countries and respective technological similarities between countries as vertices and edges. Applying this framework, the research questions are addressed qualitatively by identifying the structure of the network and how it has evolved. The analysis results indicate that the global knowledge network consists of a cluster of developed countries, and the cluster is linked with developing countries through Japan, U.S.A. and China. They also show that the Information and Communication Technology leaders changed from Great Britain and France to U.S.A. in 1920s, from U.S.A. to Japan in 1970s. The framework is expected to be applied to economic studies of innovation and knowledge bases at a global level.

Cloud computing breaks up the traditional value chain of IT provisioning and leads to new roles of market players acting in the ecosystem. Although there exist few publications on modeling the cloud computing ecosystem, each contains a different number and various types of roles. The goal of this research paper is, therefore, to perform a comparative analysis of the dominating cloud computing ecosystem models in order to develop a revised, more comprehensive model. After having excluded several roles assessed as being irrelevant and included the findings of eight interviews with experts from cloud computing service providers, the Passau Cloud Computing Ecosystem Model (PaCE Model) comprises 18 roles. This model serves as a basis to investigate whether each role can actually be covered by real actors and which typical role clusters prevail in practice. Practitioners can gain a deeper understanding of the ecosystem’s complexity and recognize where they are situated and how they are related to each other.