Bilateral research and industrial development enhancing and integrating GRID Enabled Technologies

Informazioni relative al progetto

BRIDGE

ID dell’accordo di sovvenzione: 045609

Progetto chiuso

Data di avvio 1 Gennaio 2007

Data di completamento 31 Dicembre 2008

Finanziato da

• Information Society Technologies: thematic priority under the specific programme "Integrating and strengthening the European research area" (2002-2006).

Costo totale € 2 481 350,00

Contributo UE € 1 699 999,00

Coordinato da

FRAUNHOFER IAF
Germany

Exploitable results

Europe links with China for twice the computational power

Bridge was a Specific targeted research project (STREP) focussing on demonstrating the benefits of grid technology for international cooperation, in particular between Europe and the target country China. The project has been a success, with many results already in commercial applications. The Bridge strategic objectives and the related outcome are: - To demonstrate the benefits of grid technology for international cooperation Bridge has been highly successful. IT Innovation and BUAA have developed a grid infrastructure. First the requirements were analysed before specifications were implemented by GRIA and extensions to provide the performance required by the Bridge applications and interoperability with GOS. The differences between GRIA and GOS were handled through a series of phased developments, starting with a simple but limited gateway-based approach to interoperability, which was then replaced by a more generic client-side API capable of invoking different types of service. This was then augmented by an 'adaptive API', which provides a common interface to workflow authors. Finally, IT Innovation and BUAA looked at some newer technology stacks and began the process of migration to one of them in order to take advantage of the newer standards support that these platforms offer. - To develop, enhance and interconnect European and Chinese grid middleware technology GRIA and GOS do have significant similarities in core message level protocols and the services provided for data processing and storage but there are important differences in approaches to management of business relationships. Bridge achieved interoperability by federating services using a combination of adaptive client interfaces (disparate approaches) and services based on standardised interfaces (analogous approaches). Three main interoperability areas were considered: 1. Policy management: initiating trust relationships, managing security policies, and maintaining access rights process state. 2. Resource and capacity management: deciding how to provision service considering service providers resource constraints. 3. Data processing and storage: executing computational jobs, requesting data transfer and storage. - To set up integrated grid test bed using European and Chinese middleware components for application demonstration GRIA and GOS both supported a number of these standards before the Bridge project. IT Innovation and BUAA have together been able to deliver a powerful and consistent infrastructure which has allowed the Bridge application partners to collaborate effectively in running integrated application workflows that utilise grid resources across Europe and China. - To disseminate the results of the project to industrial and academic communities Bridge results have been presented on a large number of conferences, both in China and Europe. Especially the Chinese partners have organised and hosted many workshops, where industrial partners from the relevant Chinese industrial sectors participated. In addition, special application sector specific events were organised targeting scientists and engineers from these application areas. Project flyers and press releases have promoted the results and findings of Bridge to a wider audience and were widely taken up. - To provide a software platform supporting distributed product and process developments, which respects and protects intellectual property rights Bridge has developed a service-oriented interoperability framework between European and Chinese middleware technologies that allows the execution of distributed workflows with transparent access to distributed data repositories and heterogeneous grid resources. European sites provide services based on GRIA and Chinese sites to provide services based on Cngrid GOS. Interoperability focuses on supporting secure and managed distributed collaborations between Chinese and European industrial and academic communities, whilst satisfying the specific needs of the target application sectors. - To set up joint application show cases using distributed workflow and data access technology Demanding applications show cases have been defined in each of the three application sectors, which have been solved be the project partners implementing related prototypes: Aerospace The aerospace scenario has been for the Bridge project a though challenge to accomplish. Starting from the original objectives of the DOW, the complexity of the implementation of this scenario has been unveiled and its representativeness of industrial aerospace applications has been clearly demonstrated in the two prototypes. Meteorology During the course of the Bridge project, the partners of the meteorological activity have deployed a hybrid grid infrastructure based on GRIA in Europe and GOS in China. The grid offers access to model outputs as well as basic analysis services, such as the computation of averages, standard deviation and plotting facilities amongst many others. Using the Metview macro languages, a power full language used in the meteorological community to manipulate model outputs, users can chain these services to compute elaborated probabilistic products, such as clusters or Epsgrams, in a distributed fashion over the grid. Pharmaceutical In Dockglow the Pharma partners have successfully integrated the four different docking tools executing as remote services at geographically distributed locations in Europe and China from a common workflow. The Pharma partners used multiple algorithms for solving the same problem in a branched workflow to address the problem of diversity of results produced by different docking tools. The storage of the outcomes of these tools in a common docking database supports a comparison of the results. Moreover, the Pharma partners could demonstrate on a test set that our combinatorial approach Combiscore improves docking accuracy compared to the use of a single tool. Environments like the DDB support result analysis in such large scale collaborative screening experiments by providing a common platform with sorting, filtering and visualizing facilities but the current implementation of the DDB is not suited to be easily integrated as a Grid service in a generic way with other tools, and has been specifically integrated for the Dockflow prototype. This clearly marks a limitation of virtualisation with the tools and techniques currently available. There is a long road to go for software tools routinely used in cheminformatics to enable them as services integrated in a grid environment. Our next step in building a truly integrated problem-solving environment for virtual screening would aim at an integration of the docking database (DDB) as a true grid resource. As on the technical site implementations of the core grid technology already exist, like in the Oracle 10g family of software products, we do not anticipate major problems at this point. Once we have the DDB integrated as a true grid service, the prototypic Pharmagrid will be turned into a truly productive environment that supports virtual screening with the ability to optimise docking strategies and to make full use of the adaptive features of the workflow.