The main goal of Sirene is to provide a better infrastructure by exchanging resources. For instance, some countries have very large supercomputers, others have specialized visualisation and virtual reality expertise. By connecting these in a seamless Grid, the countries can still do what they are best at, but provide the other services through the resources of partnering countries.
What is important is to provide and develop good user interfaces, especially for scientific disciplines that are new to the Grid and not technical. The current interfaces seem to be developed by techical people for technical people, according to Patrick Aerts, co-chair of Sirene. There should come a no-barrieres intuitive entry into a resource on the Grid for any user. Advanced high speed networking will provide the basis for the infrastructure. In Grid middleware, Sirene will also look for standard solutions, to allow coupling of the national Grids. It will also follow standards implemented worldwide and elsewhere in Europe so the Sirene part can be part of the worldwide Grid.
Sirene focuses on Grid infrastructure for eScience, not commercial applications. eInfrastructures to support science are flourishing and developing quickly. Hence it makes sense to focus on that. Pierre Bruyère from the Belgium networking organisation BELnet that is also operating the national BEGrid was illustrating this. BEGrid currently has about 500 processors, 4 Tbyte of storage and about 400 users. Because of the size of the country, the Grid will never grow to the size of that of, for instance, Germany or France, but a lot of issues are the same: legal, security, to mention a few. Hence it is useful to co-operate with similar organisations and share the knowledge.
Patrick Aerts stressed that Sirene is not another Grid project. In fact it is not a project at all. It is a collaboration and the signing of the memorandum shows the political will to work closely together and make it a success. The Sirene focus is on the development of the second generation of Grids: co-operations with projects that are developing Grid infrastructures on a European level, such as EGEE and DEISA. General Grid eInfrastructure policies on a European scale are developed within the eInfrastructures reflection group (e-IRG). Of course, Sirene will follow the recommendations of e-IRG as much as possible.
Sirene will take a pragmatic view. Hence, as soon as possible some major resources will be made available and an inventory with resources suitable for sharing will be produced. For major activities, like the coupling of Grids, Sirene will seek funding.
What are the tangible results we can expect from Sirene? Patrick Aerts pointed at the new supercomputers that will be available shortly in Finland and The Netherlands. Making them available is a first visible result. With, for instance, a new professor in man-machine interfaces at Delft University, discussions have been started on how, for instance, scholars in human sciences should view a new resource like a Grid.
In Switzerland, for instance, the financial research community is interested in using the Grid. They have completely different approaches to computers and how to use them than the more traditional physics and chemistry disciplines.
There are already several nation-crossing Grid initiatives, like the Nordic Grid and the Baltic Grid. What does Sirene add to that? According to Valdas Rapsevicius, a lot. The scale of Sirene is larger, and the range of expertise broader. Hence in Lithuania, one expects to benefit from the broad knowledge base within Sirene. They also have installed a supercomputer recently, but lack applications to support users. Hence, exchanging supercomputer time for application knowledge seems an approach that could be taken within Sirene. Sirene could give a boost to science in individual countries.
In Switzerland one expects to learn from the different strengths to serve different types of users. In some countries physicists are the dominant Grid user group, in others not.
Why is Sirene only focusing on supporting scientific use of the Grid? The main reason is that the organisations that signed the memory of understanding are supporting eScience in their countries. The funding models, although they may differ from country to country, are fundamentally different from the ones used by commercial co-operations. Also the local eInfrastructure is intended to be used by scientific researchers. Mixing that with commercial usage is difficult. But the main reason is that in science there is a need for international collaboration. Data sources and databases, for instance, are often produced by international collaborations and when using them, they are combined with, for instance, supercomputing or visualisation resources.
Will Sirene provide a roadmap of what they intend to do? That is not the main priority. There is enough practical work to do and gaining experience in linking the Grids and exchanging resources. More information can be found at the Sirene web site.