Antonella Galizia

Hybrid Clouds brokering: Business opportunities, QoS and energy-saving issues

Abstract Hybrid Clouds couple the scalability offered by public Clouds with the greater control supplied by private ones. A (hybrid) Cloud broker acting as an intermediary between users and providers of public Cloud services, may support customers in the selection of the most suitable offers, optionally adding the provisioning of dedicated services with higher levels of quality. The paper presents a Cloud brokering algorithm delivering services with different level of non-functional requirements, to the private or public resources, on the basis of different scheduling criteria. With the objective of maximize user satisfaction and broker’s revenues, the algorithm pursues profit increases by reducing energy costs, through the adoption of energy saving mechanisms. A simulation model is used to evaluate performance in terms of broker’s revenue, user satisfaction and energy behavior of various allocation policies. Simulation results show that differences among policies depend on system loads and that the use of turn on and off techniques greatly improves energy savings at low and medium load rates.

Job-resource matchmaking on Grid through two-level benchmarking

Grid environments must provide effective mechanisms able to select the most adequate resources satisfying application requirements. A description of applications and resources, grounded on a common and shared basis, is crucial to favour an effective pairing. A suitable criterion to match demand with supply is to characterize resources by means of their performance evaluated through the execution of low-level and application-specific benchmarks. We present GREEN, a distributed Matchmaker, based on a two-level benchmarking methodology. GREEN facilitates the ranking of Grid resources and the submission of jobs to the Grid, through the specification of both syntactic and performance requirements, independently of the underlying middleware and thus fostering Grid interoperability.

Job allocation strategies for energy-aware and efficient Grid infrastructures

Complex distributed architectures, like Grid, supply effective platforms to solve computations on huge datasets, often at the cost of increased power consumption. This energy issue affects the sustainability of the infrastructures and increases their environmental impact. On the other hand, due to Grid heterogeneity and scalability, possible power savings could be achieved if effective energy-aware allocation policies were adopted. These policies are meant to implement a better coupling between application requirements and the Grid resources, also taking energy parameters into account. In this paper, we discuss different allocation strategies which address jobs submitted to Grid resources, subject to efficiency and energy constraints. Our aim is to analyze the potential benefits that can be obtained from the adoption of a metric able to capture both performance and energy-savings. Based on an experimental study, we simulated two alternative scenarios aimed at comparing the behavior of different strategies for allocating jobs to resources. Moreover we introduced the Performance/Energy Trade-off function as a useful means to evaluate the tendency of an allocation strategy toward efficiency or power consumption. Our conclusion seems to suggest that performance and energy-savings are not always enemies, and these objectives may be combined if suitable energy metrics are adopted.

Scheduling strategies for enabling meteorological simulation on hybrid clouds

The flexible and pay-as-you-go computing capabilities offered by Cloud infrastructures are very attractive for high-demanding e-Science applications like weather prediction simulators. For their ability to couple the scalability offered by public service provider with the greater control and customization provided by Private Clouds, Hybrid Clouds seem a particularly appealing solution to support meteorological researchers and weather departments in their every-day activity. Cloud Brokers interfacing customers with Cloud providers, may support scientists in the deployment and execution of demanding meteorological simulations, by hiding all the intricacies related to the management of powerful but often complex HPC systems.The paper presents a set of brokering strategies for Hybrid Clouds aimed at the execution of various instances of the weather prediction WRF model subject to different user requirements and computational conditions. A simulation-based analysis documents the performance of the different scheduling strategies at varying workloads and system configuration.

A QoS-aware broker for hybrid clouds

Hybrid Clouds seems able to offer their customers with differentiate solutions capable of providing more and personalized guarantees with respect to the basic service availability generally supplied. In the context of an Italian research project aimed to transfer ICT advancements from research centers towards ICT SMEs, the paper focuses on the design of a brokering tool for hybrid clouds capable to adequately respond to specific Quality of Service (QoS) constraints. Aimed at satisfying the highest number of user requests while trying maximizing the profit of the private provider, in the context of a posted price economic model, the proposed brokering algorithm may apply different allocation policies, based on the reservation of a quota of private resources to high-level QoS applications.

An object interface for interoperability of image processing parallel library in a distributed environment

Image processing applications are computing demanding and since a long time much attention has been paid to the use of parallel processing. Emerging distributed and Grid based architectures represent new and well suited platforms that promise the availability of the required computational power. In this direction image processing has to evolve to heterogeneous environments, and a crucial aspect is represented by the interoperability and reuse of available and high performance code. This paper describes our experience in the development of PIMA(GE)2, Parallel IMAGE processing GEnoa server, obtained wrapping a library using the CORBA framework. Our aim is to obtain a high level of flexibility and dynamicity in the server architecture with a possible limited overhead. The design of a hierarchy of image processing operation objects and the development of the server interface are discussed.

Lessons learned implementing a science gateway for hydro-meteorological research

A full hydrometeorological (HM) simulation, from rainfall to impact on urban areas, is a multidisciplinary job, which relies on the execution of a workflow composed of complex and heterogeneous model engines. Moreover, the accuracy of the simulation is strongly dependent on an extensive set of configuration parameters, which have to be selected in a consistent way among the models. Within the Distributed Research Infrastructure for Hydro‐Meteorology project, a Web‐based science gateway was developed with the aim to support HM researchers in designing, executing, and managing HM experiments. The core of this science gateway is the portal, which takes care of generating all the configuration files and handles the execution of simulation steps on a heterogeneous computing infrastructure composed of high‐performance computing, Grid resources, and Cloud resources. This paper presents technological insights about the implementation of the portal, with an analysis of the adopted technologies and infrastructures. Our experience highlights the need of coherent policies in the management of data, computational resources, and software components that represent the ecosystem to develop science gateways. Copyright

An MPI-CUDA library for image processing on HPC architectures

Scientific image processing is a topic of interest for a broad scientific community since it is a mean of gaining understanding and insight into the data for a growing number of applications. Furthermore, the technological evolution permits large data acquisition, with sophisticated instruments, and their elaboration through complex multidisciplinary applications, resulting in datasets that are growing at an extremely rapid pace. This results in the need of huge computational power for the processing. It is necessary to move towards High Performance Computing (HPC) and to develop proper parallel implementations of image processing algorithms/operations. Modern HPC resources are typically highly heterogeneous systems, composed of multiple CPUs and accelerators such as Graphics Processing Units (GPUs) and Field-Programmable Gate Arrays (FPGAs). The actual barrier posed by heterogeneous HPC resources is the development and/or the performance efficient porting of software on such complex architectures. In this context, the aim of this work is to enable image processing on cluster of GPUs, through the use of PIMA(GE)2 Lib, the Parallel IMAGE processing GEnoa Library. The library is able to exploit traditional clusters through MPI, GPU device through CUDA and a first experimentation is aimed to explore the use of GPU-clusters. Library operations are provided to the users through a sequential interface defined to hide the parallelism of the computation. The parallel computation, at each level, is managed employing specific policies designed to suitably coordinate the parallel processes/threads involved in the elaboration and their use is tightly coupled with the PIMA(GE)2 Lib interface. In this paper, we present the incremental approach adopted in the development of the library and the performance gains in each implementations: quite linear speedup is achieved on cluster architecture, about a 30% improvement in the execution time on a single GPU and the first results on cluster of GPUs are promising.

Setting Up an Hydro-Meteo Experiment in Minutes: The DRIHM e-Infrastructure for HM Research

Predicting weather and climate and its impacts on the environment, including hazards such as floods and landslides, is a big challenge that can be efficiently supported by a distributed and heterogeneous infrastructure, exploiting several kinds of computational resources: HPC, Grids and Clouds. This can help researchers in speeding up experiments, improve resolution and accuracy, simulate with different numerical models and model chains. Such numerical models are complex with heavy computational requirements, huge numbers of parameters to tune, and not fully standardized interfaces. Hence, each research entity is usually focusing on a limited set of tools and hard-wired solutions to enable their interaction. The DRIHM approach is based on strong standardization, well defined interfaces, and an easy to use web interface for model configuration and experiment definition. A researcher can easily compare outputs from different hydrologic models forced by the same meteorological model, or compare different meteorological models to validate or improve her research. This paper presents the benefit of a web-based interface for hydro-meteorology research through a detailed analysis of the portal (based on liferay-gUse) developed by the DRIHM project.

CUDA accelerated blobby molecular surface generation

A proper and efficient representation of molecular surfaces is an important issue in biophysics from several view points. Molecular surfaces indeed are used for different aims, in particular for visualization, as support tools for biologists, computation, in electrostatics problems involving implicit solvents (e.g. while solving the Poisson-Boltzmann equation) or for molecular dynamics simulations. This problem has been recognized in the literature, resulting in a multitude of algorithms that differ on the basis of the adopted representation and the approach/ technology used. Among several molecular surface definitions, the Blobby surface is particularly appealing from the computational and the graphics point of view. In the paper we describe an efficient software component able to produce high-resolution Blobby surfaces for very large molecules using the CUDA architecture. Experimental results show a speedup of 35.4 considering a molecule of 90,898 atoms and a resulting mesh of 168 million triangles.