Exploitable results
The Grid4All project was formulated almost 4 years ago. At that time, grids represented the state of the art in general-purpose distributed systems, and had some of the features required for democratic scenarios, such as sharing and aggregation of resources, and forming Virtual Organisations (VOs) across organisational boundaries. Our vision was naturally expressed as a generalisation of grids, a movement from Grid-For-Some to Grid4All. Today, an appropriate title might be Hybrid, Democratic Cloud.
The project was set up to enable open, democratic access to computing resources, supporting both grid and edge resources, and to support co-operation within distributed groups of users, so-called virtual organisations. The name Grid4All, and its slogan democratic Grid, highlight its focus on supporting small users, having limited financial, physical, and system administration resources. To these users, important factors are openness, the ability to utilise all available resources (including volatile ones), ease of use and self management. The original project objectives may be divided in two themes (a) pooling resources and utilising them effectively and (b) collaboration and sharing data.
To achieve its objectives, the project developed prototype implementations and the execution of the project was guided by the following principles:
Architecture-based management approach:
Grid4All follows the architecture-based management approach, which uses architectural models as the basis of deployment, monitoring and reconfiguration activities.
Integration based on innovative scenarios:
Grid4All has designed and developed a set of software modules that commonly address the overall project objectives. While complete integration of all software modules is clearly infeasible, design and development have followed common architectural principles. Integration of different components provides enhanced capabilities; e.g. VOFS integrated to the Grid4All security architecture enforces access control on shared workspaces. The project has identified and implemented the most innovative integrative scenarios.
Prototyping of new applications and adaptation of existing applications
The project aims to enable "democratic grids", i.e. to provide new classes of users with flexible access to large computational capacities, to enable them to pool their computational resources over the network, and to empower them with sharing and collaboration facilities. Middleware, services and tools are designed to work around the problems intrinsic to such grids (volatility, churn, heterogeneity etc) and to users of such grids (different levels of IT awareness and skills, disconnected mode of work etc). Applications are not necessarily new in terms of functionality, but need to execute and provide services even when deployed in such environments and for and by such users. To evaluate Grid4All technology results, we need to not only prototype new applications but also adapt (as minimally as possible) existing applications.
Evaluation:
Qualitative evaluation procedures that the project had defined allowed obtaining the perspective from end-users and their appreciation of the functional and non-functional aspects of the system. We had identified two categories of evaluators: end-users who install, set-up and use the software product and developers who use the product (middleware) to design and develop new applications.
Case studies:
The implementation of the architecture will be put into practice in real-world case study, serving both as a proof of concept for the architecture in all of its scenarios, and also as a vehicle to analyze empirical usage findings in a real-world context.
Management Tools
Managing distributed applications operating on large-scale, dynamic environments, such as democratic grids or clouds, is complex and error-prone. Application management generally consists of two parts. First, there is the initial deployment and configuration, where individual components are shipped, deployed, and initialized at suitable nodes, then the components are bound to each other as dictated by the application architecture, and the application can start working. Second, there is dynamic reconfiguration where a running application needs to be reconfigured. To reduce the complexity of those tasks, Grid4All proposes a set of management tools based on the component-based management approach; the tools include the Niche platform, an ADL-based deployment service, and the DepOz framework.
Applications and scenarios
End users perceive the usefulness of the platform indirectly by the functional and non-functional behaviour of applications and the environment within which the applications are capable of executing. It is clear that most applications that are adopted by the project do already exist, but these currently execute within controlled environments - expensive infrastructures, high level of administration, expensive software. To demonstrate the feasibility of the Grid4All software platform in addressing requirements of a democratic Grid we have designed and developed groupware applications using Grid4All technology enablers in different application prototypes. Usage scenarios derive requirements for infrastructure and applications for Grid4All intended uses. While some applications are designed from scratch, others are legacy applications that have been redesigned and adapted. This has helped understand the complexity imposed by dynamic environments on applications and define algorithms to overcome situations and limitations. Adhering to our objectives, we have selected a range of applications of interest to end-user focus groups targeted by Grid4All. Collaborative File Sharing, eMeeting and Shared Calendar applications support collaborations; collaborative work is essential for teaching and research and for non-profit organisations, leisure and social networks.
Searching for OpenAIRE data...
There was an error trying to search data from OpenAIRE
No results available