Abstract: GaiaGrid was established to answer three main objectives:
The software platform that was used to respond to these objectives was the a standard Globus Toolkit 2 environment in conjunction with a Workflow Tool developed by DutchSpace: GridAssist. Today, GaiaGrid is composed of about 40 CPUs scattered across nine institutes in seven countries. The results of these initial tests have been very promising and have shown that with little budgetary investments (most of the machines used in the ESTEC Grid for instance were old PC's that were recycled) excellent performance results could be achieved. This presentation will discuss some of these resutls in detail and highlight some of the cost benefits in this area.
Abstract: Building and operating computing and storage infrastructures based on grid technology is a difficult task. There is little reference on providing information technology services based on a grid infrastructure at the global level. One of the main goals of the european project EGEE (Enabling Grids for E-SciencE) is to build a production-level infrastructure based on grid technology for the scientific community. As one of the sites contributing to the global operations of the EGEE infrastructure, we will present what we have learned during the first year of the project on this subject and what issues need to be addressed when building an international grid.
Abstract: The German Government is funding a nationwide effort to make Grid-Technology available to scientific Communities. AstroGrid-D is one of the Community-Projects acting as 'Science-Driver' within D-Grid.
The AstroGrid-D has, asides developping a sound grid-infrastructure for German Astronomy, three goals:
Although working within a national context (D-Grid: www.d-grid.de), AstroGrid-D (www.gac-grid.org) will collaborate with grid projects of astronomy communities across national boundaries. We keep close relations to IVOA, supporting efforts for worldwide interoperability and standardization of access to astronomical data.
Abstract: Today, spatial data processing centers are usually organized in a centralized workflow of operations as acquisition, archiving, processing and distribution functions. Scientific laboratories achieve processing for level 1, 2, 3 products while CNES and other spatial agencies are responsible for the system architecture. In case of future ground segments which require huge storage and processing capabilities, the Grid concept can be an opportunity to deploy operational distributed ground segments through Europe (and beyond) for which CNES can gather and share its computational and storage resources. That is why CNES has started a new study with CS company in order to examine the grid technology and its potential usage in data processing centers where actors and processing resources are geographically distributed. This study consists also in gridifying a specific image processing application (Chistera for Spot5 satellite ground segment) and in studying security constraints in order to build an extra grid where CNES (Toulouse) will be involved. The CNES presentation will be focused on the security study, its first technical results in term of architecture definition (over Globus TK4) and CNES experience feedbacks.
Abstract: Earth scientists use large amounts of electronicly available data, information and associated knowledge from different satellite sensors, aerial campaigns and in-situ measurements in combination with related models, tools and documents to perform their research. Accessing the right data and requested associated information to understand and use them isn't straightforward however.
Scientist may benefit in their struggle for data by adapting more e-collaborative ways of doing science also known as e-Science.
The presentation provides an outline of how different e-collaboration technologies, including Grid and web-services, digital library and more, can help in improving Earth science data access by describing progress in the European Space Agency (ESA).
In particular it will be presented progress in:
Recent References:
Abstract: Since the last decade, the technological developments for the computing grids were focused on storage, computing and resources management in order to create a very powerful environment for the management and the analysis of scientific data. Therefore, the aim of a grid is to place at the disposal of the users a geographically distributed collaborative environment which enables them to achieve more efficiently their objectives.
RUGBI and Openplast are industrial and academic projects funded respectively by the French Ministry of Research (Réseau GenHomme Action Génomique et Innovations Médicales) and RNTL (Réseau national de recherche et d'innovation en Technologies logicielles). These computing and data grids are designed and deployed on top of existing technologies to offer a set of services to analyse proteins for RUGBI and plastic molding or injection simulation for Openplast. They aim to support SMEs for computing, storage and deploy interregional grids for bioinformatics or plasturgical business.
These grids offer specific business oriented tools through a user-friendly portal in a secure, confidential-aware and industrial environment. Performance, security and transparency of the web portal are enhanced by the distributed, interoperable and close to the portal grid information system (workflow, access controls, ...). Based on existing grid technology such as Globus Toolkit 4.0 and OGSI, reusable and security-aware (confidentiality and integrity) grid middleware components has been developed within the framework of theses projects. Innovative components for managing the coordination of various tasks types (jobs, transfers of data sets, service operation invocation, ...) and for aggregating, structuring and filtering log events, are examples of services developed in this project.
Abstract: Parallelism has successfully reduced the time requested for running CFD computations, and allowed computation on bigger meshes to capture more precisely the aerodynamic phenomenons. This has led to typical industrial use cases that defeat the usual batch queuing management efficiency, and request smarter scheduling that only grid strategies can implement. Two typical industrial use cases will be presented, namely using HPC ressources for interactive visualization without dedicating them to interactive work, and efficiently interleaving optimization or multi-disciplinary simulations that are only partly parallel to avoid leaving parallel ressources idle while sequential parts are executed. A typical solving strategy will be presented, that can be implemented on several existing grid solutions.
Abstract: The GRID-TLSE project funded by ACI-GRID aims at developing a web expert site in sparse linear algebra.
During this talk we will focus on the main difficulties that arise when deploying and managing a large amount of scientific software over a grid.
We will then describe the goals and the status of the GRID'5000 nation-wide experimental plaform for grid research with special consideration on the research projects developped within "Grid'mip": the Toulouse / Midi-Pyrénées Site of GRID'5000.
The EADS Corporate Research Centre is currently constructing an internal Grid infrastructure - in between its different sites (Suresnes, Toulouse in France and Ottobrun in Germany). This grid is designed for our internal applications but also as an in-house "laboratory" in which we can exercise grid concepts, capabilities and tools to develop and evaluate a range of Grid enabled business scenarios. This is facilitated by our close relationship with world-leading French and EU projects. Grid technology in its current form is ready to be deployed in areas of compute-intensive activity and can demonstrably improve efficiency of deployed IT assets within a single organisational domain. However, the big pay-off from Grid is in the collaborative "Virtual Organisation" area. This talk will present the current EADS application requirements and give an overview of the different commercial and open-source frameworks that have been evaluated.
(This presentation is not yet available online)