supported by the EC IST Programme
CoreGRID European Research Network on Foundations, Software Infrastructures and Applications
for large scale distributed, GRID and Peer-to-Peer Technologies
image
Home arrow CoreGRID NoE arrow Objectives  

spacer spacer
 
Main Menu
Home
News
Events
CoreGRID WG
CoreGRID NoE
Institutes
Integration Activities
Dissemination
Training & Education
CoreGRID & Industry
Mobility Portal
Trust&Security Portal
Collaboration Gateway
Other Collaborations
Links
 
spacer spacer
spacer spacer
 
CoreGRID Objectives

GRID computing emerged in the mid-nineties as a new high-performance computing infrastructure for scientific and engineering applications. The GRID (Foster & Kesselman 1998) focussed in its early stage mainly on metacomputing: sharing computing resources over the Internet. The current architectural extensions based on the GRID are not sufficient to meet the broader requirements such as those required for business or heavy-duty scientific needs. Such requirements include scalability, robustness, dynamicity, self-healing, high integrity, business-strength security and trust, low effort threshold end-user interface, homogeneous access to heterogeneous resources and sources. The GRID concept, the Network is targeting, is therefore a much broader vision than the GRID as it is today and it represents a challenge for the next ten years. Our vision of the Grid will allow an end-user to interact with the system to specify a request and the system will respond with a 'deal' that the user may accept or reject. The deal will encompass different kinds of resources such as data, information and knowledge, sensors or particular computing equipments such as visualisation systems as well as computational resources. Our vision of the GRID will be based on a number of layered components (see Figure 1), some of which are now well understood and others, which are currently only speculative. All of them, however, require further R&D and improvement.

vision.gif

Figure 1: CoreGRID vision of the Next Generation GRID

CoreGRID aims at building a virtual European-wide Research Laboratory that will achieve scientific and technological excellence in the domain of large scale distributed, GRID and Peer-to-Peer technologies aiming at realizing our vision of what should be a GRID infrastructure. This vision of a new generation GRIDs architecture integrates in a seamless way the existing GRID and other emerging architectures such as P2P (peer-to-peer) taking in W3C (world-wide-web consortium) and other relevant standardization bodies concepts and standards.

It is the primary objective of the CoreGRID Network of Excellence to build solid foundations for GRID and Peer-to-Peer on both a methodological basis and a technological basis and to stay at the forefront of the Excellence. This will be achieved by structuring research activities, leading to integrated research among experts from the relevant fields, and more specifically distributed systems and middleware, programming models, algorithms and tools and environments. Our long-term vision of the GRID (10 years) requires that CoreGRID must be sustainable and must evolve by itself when European funding will cease.

To comply with this long-term objective, the CoreGRID Network defines and conducts a joint programme of activities (JPA) in order to integrate and co-ordinate the ongoing research activities of the major European research teams in the field of Grid and P2P technologies for the purpose of developing the Next Generation GRID. These research teams bring both their high-level expertises in the particular areas and also their own capabilities in influencing their national programmes in Grid and P2P technologies allowing thus a better long-term integration (in the sense given by the EC). CoreGRID participants are involved in 15 Grid-related national programmes within which they have a key role (for example, members of the scientific committees, programme directors, etc.) in the definition of the research agenda.

The implementation of an integrated programme of this type, as it is briefly described in this section, will make the accomplishment of our grand vision for Next Generation GRID feasible. This programme has two main objectives: excellence and integration. An ambitious programme for jointly executed research activities by the best teams in Europe and a set of integrated activities will be implemented to reach these two objectives.

A programme for jointly executed research activities to promote excellence

The research programme that will be undertaken within CoreGRID is structured around six complementary research areas as listed below. These areas have been selected on the basis of their strategic importance, their research challenges and the ability to gather at least three different research teams from three different countries. This latter constraint was imposed to ensure a real integration of the research carried out in Europe and to fit the goal of a Network of Excellence as it is defined by the EC in FP6. The six research areas are as follows:

  • Knowledge & Data Management: This area will focus on information and knowledge GRIDs. More specifically this area will study various techniques and tools for supporting data intensive applications and the integration of data and computational GRIDs with information and knowledge GRIDs. Knowledge GRIDs offer high-level tools and techniques for the distributed mining and extraction of knowledge from data repositories available on the GRID. The development of techniques, tools and services for such an infrastructure is the main goal of these research activities, focused on the design and implementation of an environment for geographically distributed high-performance knowledge discovery applications. The main objective of the network in this area is to provide a collaborative infrastructure for European research teams working on distributed storage management on GRIDs, development techniques and tools for supporting data intensive applications and integration of data and computational GRIDs with information and knowledge GRIDs. These efforts represent a significant step in the process of studying the unification of data management and knowledge discovery with GRID technologies for providing knowledge-based GRID services as required for the European knowledge-based society.
  • Programming Model: This area will investigate the definition of programming models for the GRID by reducing the complexity of programming GRIDs. In particular, a software component model for GRIDs will be researched to provide a higher level of abstraction compared to the current practices mostly based on traditional message-passing primitives. New GRID programming models are necessary that rely on a higher level of abstraction and are based on component technology. The main objective in this area is to define a novel common European component model, suitable for future large-scale GRID computing and P2P. The new component model will meet the main challenges of Next Generation GRID systems and their applications. In order to deal with scale and heterogeneity, the component model must be hierarchical, and to enhance programmability the component model should allow the use of structured composition to design new components. To deal with dynamicity, it must be capable of handling reconfiguration. Moreover, the component model should be able to handle components at the infrastructure level in order to benefit from predefined, GRID-aware, efficient, composable and reusable service components.
  • Architectural Issues: Scalability, Dependability, Adaptability : This area will cover the study of distributed system architecture for next generation GRIDs with particular emphasis on scalable GRID services, adaptability and dependability. Next generation GRID systems may need to scale to millions (if not billions) of nodes. Such an increase in scale may also require fundamental changes in the underlying GRID architecture. Regarding dependability, one of the main challenges for GRID computing is the ability to tolerate failures and recover from them (ideally in a transparent way), since current GRID middleware still lacks mature fault tolerant features. Based on the above observations, the main objective on this research area is to investigate scalable approaches in the design of next generation GRID as well as to improve the robustness and dependability of GRID services and middleware. The development of mechanisms for automated adaptation and reconfiguration of the GRIDs on all hierarchy levels will be carried out. Such mechanisms will include monitoring of the state of the GRID, its analysis together with decision taking, intelligent decision execution and finally delegation of control to human operators.
  • Grid Information, Resource and Workflow Monitoring Services: Next Generation GRIDs require coordination of several core services to achieve useful integration of resources provided by different resource owners. This is impossible to achieve without relevant information about the state of resources and services, properly collected, merged and filtered if necessary. Current information and monitoring frameworks do not scale to GRID level or are focused on specific aspects. Also, a generally acceptable set of metrics to evaluate the GRID performance is still not available. The primary objective in this research is the development of a general and scalable information and monitoring infrastructure that will cover both service and job centric views of GRID. New GRID performance models must become available to provide foundation for means and tools for the evaluation of services deployed on the GRID. Methods for the complex job workflow extraction from programming models, their monitoring, and execution support with adequate checkpointing and migration support must are also necessary. All together this will provide the necessary framework and information base for the underlying resource management and scheduling as required by a "Next Generation GRID".
  • Resource Management and Scheduling: The management and coordination of resources in a GRID environment is a key topic for future GRIDs, which has also been identified by the NGG expert group. Currently, there is neither a coherent and generally accepted infrastructure to manage and schedule resources nor are there efficient coordination algorithms that suit the complex requirements of a large scale GRID environment with different resource types. Key issues in this area are, for instance, the consideration of a cost and accounting model as well as the support for individual policies of GRID users and resource providers. Therefore, the main objective in this research area is the development of a common and generic solution for GRID scheduling and management in Next Generation Grids. This includes the architectural perspective which has to support scalability and cooperation across different administrative domains. Such an architecture requires new scheduling algorithms and scheduling policies for coordinating the access to resources and for supporting complex job requirements. The actual management of resources and jobs further needs several services that provide the necessary functionality e.g. to manage the scheduling of complex job workflows
  • Grid Systems, Tools and Environments: This area concerns the design of a software environment and tools for GRID systems. The main objective is to build a generic component system that integrates application components, tools/system components, problem solving environments (PSE)/portal components, and infrastructure components. Our adopted approach encompasses both the realization of the component model, and its integration into a tools framework. It will be composed of an integrated components GRID platform, a component support toolkit and a generic problem-solving toolkit. The design of such a platform will advance significantly the state-of-the-art. It will enable the building of a single component-based platform for both applications and tools/systems/PSEs to have a single, seamless and "invisible" GRID software infrastructure. The results from this research are expected to be equally valid for both the client/server and the peer-to-peer paradigms.
  • Reference: Ian Foster & Carl Kesselman (1998). The Grid: Blueprint for a New Computing Infrastructure. Morgan Kaufmann Publishers, San Francisco, California.

Last Updated ( Friday, 23 March 2007 )
 
 
spacer spacer
spacer spacer
 
© 2012 CoreGRID Network of Excellence - European Grid Research
Joomla! is Free Software released under the GNU/GPL License.
 
spacer spacer