Search

Efficient resource management is crucial for balancing performance and energy consumption in large-scale data centres. In the case of additional requirements such as guaranteed resources and low communication latency, it is of great importance to implement not only an efficient initial placement algorithm, but also maximise consolidation by migration techniques, making sure that network performance is not sacrificed. In this paper, we introduce a hierarchical approach to migrations based on a combination of efficient packing algorithms and network communities. Results analysis shows the benefits of using a two-level approach where the combination of localised consolidation and network awareness improves both performance and energy efficiency, while maintaining low network hop distance.

Hands-on experience and training related to the latest cyberthreats and best practices, augmented with real-life examples and scenarios is very important for aspiring cybersecurity specialists and IT professionals in general. However, this is not always possible either because of time, financial or technological constraints. For cybersecurity exercises to be effective they must be well prepared, the necessary equipment installed, and an appropriate level of isolation configured, preventing inter-user interference, and protecting the integrity of the platform itself. In recent years there have been numerous cybersecurity training systems developed that aim to solve these problems. They can either be used as cloud or self-hosted applications. These solutions vary in their level of sophistication and ease-of-use, but they all share a single goal, to better educate the cyber community about the most common vulnerabilities and how to overcome them. The aim of this paper is to survey and analyze popular cybersecurity training systems currently available, and to offer a taxonomy which would aid in their classification and help crystalize their possibilities and limitations, thus supporting the decision-making process.

Providing realistic production environment for deployment of complex students' projects can be quite a challenge for the university computing centers. While alternatives such as using public/private clouds exist, enabling the students to focus on core project deployment instead of system administration of a production environment in general requires an alternative approach. The proposed platform as a service solution relates to the three most important requirements of such a system: scalability, ease of use, user isolation and security.

When moving cloud computing towards the edge in order to provide low latency and location-based Internet services, there are additional considerations that need to be introduced in the traditional cloud approaches to adapt them to fog computing. This paper discusses a proposal for resource allocation and migration strategies adapted specifically to fog computing. The proposal is based on the concept of virtual and physical communities and it has already shown its efficiency when applied to cloud computing. Simulation-based scenarios provide the performance of our community-based proposal following the geographical trajectory of mobile users.

The ongoing trends in cloud computing demonstrate increasing need for efficient, yet economical data centers. Thus, recently the research community has focused its efforts on frameworks for optimised usage of the available resources that will result with energy-efficient and highly effective data centers. Toward this goal, in this paper we present a community-based framework for virtual machine placement inside a cloud data center. The framework is based on the complex network structural property of grouping tightly coupled nodes, and a matching process that maps virtual to physical communities while employing different optimisation functions on different hierarchy levels. The presented simulation results of the framework application reflect its high usage potential achieved by improvement in communication efficiency and reduced power consumption compared to the traditional heuristics.

New Grid and cloud solutions for distributed data mining and data processing are needed for execution of data intensive workflows. In contrast of the standard workflows, in which data between the jobs are exchanged in the form of files and the jobs are finished when they process the input data, data intensive workflows receive data organized in blocks which are streamed on inputs, analyze the data and produce stream output. Each job is active for a long period of time and can receive new data. In our previous research works we proposed the Open Grid Service Architecture for Data Mining (OGSA-DM), which is capable of executing data intensive workflows. According to our analysis, the current algorithms for scheduling workflows can't be applied on data intensive workflows because they produce unsatisfactory results and can't guarantee optimal solution. In this paper we propose new optimization and scheduling algorithm which is developed on the advantages of data intensive workflows. In several experiments we've shown that our proposed algorithm works and gives satisfactory results.

The problem of efficient resource management in large cloud data centres is well recognised in the cloud management community. With the rising demand for cloud services and expansion of the size and distribution of data centres, the repercussions of not taking into account all parameters of interest while allocating resources are becoming significant. Hence, one of the major issues today is how to reconcile between performance and energy efficiency so that customers can exploit the provided resources to their maximum while the provider still keeps a well balanced approach that ensures low cost and optimal power consumption. The main goal of this paper is to investigate what can be expected of performance optimisation allocation techniques from a power consumption perspective. The presented results show that careful design of resource allocation algorithms can lead towards solutions that are not only high performing, but are also good in efficient power consumption.

Modern research and education networks need to solve two major tasks: (1) providing seamless access to their users, and (2) support new scientific and collaborative applications that are becoming increasingly complex and dynamic in their scale, use of distributed resources, and required advanced networking services. Rapid deployment and automation of new network services provisioning is becoming difficult in large networks that incorporate different technologies and solutions. The task of providing seamless user experience in typically mobile and dynamically changing collaborative groups generates additional burdens for the configuration phases included in network provisioning. By introducing automation and enabling multipoint auto configuration of network devices, the Zero Touch Provisioning (ZTP) concept emerges as a possible alleviation of the complex network provisioning and infrastructure services deployment process. This paper investigates the characteristics of the ZTP model and discovers how ZTP can be used in order to enhance the services provided by the GEANT network and its associated National Research and Education Networks (NRENs) to the European research and education community. The future solution can be achieved by combining and enriching the existing provisioning models, solutions and practices available from different domains such as wireless networks, (complex) network management and operations services, and cloud based infrastructure services provisioning.

The need for customer-centric agile integrated environment offered in an on-demand self-service manner challenges the traditional R&E operations support systems. The service provider architecture proposed in this paper aims to satisfy end-user's requirements by providing a coordinated provisioning of virtualized resources through automation. Empowered by the establishment of open collaborative partnerships it offers expansion of the service catalog with multi-domain solutions based on standardized interoperability between R&E and commercial CSPs. The benefits of the approach are presented via a cloud based case study service that is transformed from a traditional silo to a business process modeled service with added values.