Feature - EGI, from the interim director's view
The European Grid Initiative, or EGI, will be one of the focal points at the EGEE'09 conference in Barcelona. In the time since EGI was announced and the plans detailed in last September's GridBriefing, it has appointed an interim director in July, Steven Newhouse, who has found the time to speak to iSGTW.
What is EGI?
Newhouse: EGI stands for the European Grid Initiative. It's a project that wil be submitted to the European Commission (EC) for funding in November. It builds on the work of the EGI-Design Study project, which looked to the grid community to identify the best models for providing a sustainable, long-term grid infrastructure to support different scientific communities within Europe.
What is the main aim of EGI?
Newhouse: The aim is to coordinate a production-quality grid infrastructure for European researchers. When grids first started up, people were sharing resources between different institutions but it was done in a very ad hoc way. The quality of the service was sometimes very low - a service you used one week might not be there the following week. But as people have adopted the grid paradigm, they have realized that in order for it to be successful these services need to be there persistently, and reliably provisioned.
The challenge is to build upon and strengthen the expertise within participating National Grid Initiatives (NGIs), so that EGI.eu - the coordinating body for EGI - can be small and focus on the coordination necessary to bring together these national grid infrastructures.
What is the difference between EGEE and EGI?
Newhouse: The operational activities within the EGI model - the infrastructure itself, the middleware and the support for the application community - will all be separate EC projects (if funded), whereas EGEE is a single integrated project. A lot of my concern as the EGI interim director is to make sure all three of these projects are going to work together and to understand the communication and management relationships between them, both at the planning stage now and also when the projects are hopefully funded in spring 2010.
Newhouse: Since September 2008 the EGI picture has become a lot clearer. The location of EGI.eu has been selected to be Amsterdam and the EC has opened up the call for distributed computing projects which can be used to support this work. The focus of the community over the last three months has shifted from the sort of visionary planning phase to the explicit "how are we going to operate EGI" phase, especially as we are now putting together the project proposals.
Where do National Grid Initiatives fit into EGI?
Newhouse: Individually, the NGI communities are coming together through the EGI Council. They have different levels of maturity in different countries, but they are all moving towards having sustainable national activities. What they are now doing is looking at how to bridge the gap between their own grid and other national grids. The role of EGI.eu is to coordinate that bridging and help enable collaboration between different countries. Many of the NGIs are already active within EGEE in various ways, and many of them want to continue and participate in the EGI project. So it certainly looks very promising at the moment.
Who will be in charge of EGI?
Newhouse: The coordinating body, EGI.eu, will be a foundation set up under Dutch law. The EGI Council, which will have representation from the NGIs and international research labs (EIROforum) which are part of the EGI collaboration, will appoint the EGI.eu board.
The EGI.eu director will manage everyday matters and work with the board to coordinate activities. The EGI council will govern the direction of the broader EGI collaboration on a longer term basis.
What will the shift to EGI mean for international scientific research?
Newhouse: One of the challenges at the moment is that some researchers have a three- or four- year applied science research project that needs to use distributed computing, but the coordination of these resources, currently EGEE, is a project only lasting two years. It's a big commitment and a risk for a scientist to use something which might disappear half-way through their project. But the EGI model should completely eliminate this risk by assuring a long-term, sustainable coordination of these resources.
So, on a day-to-day level, when we move over from EGEE to EGI, the users of the infrastructure won't see any significant changes. But longer term, as the infrastructure moves towards a more sustainable model, it will hopefully provide confidence for the research community to adopt this mode of distributed computing.
-Seth Bell, iSGTW
BONUS: Steven Newhouse talks about the Universal Middleware Distribution and standardization within EGI. Click here to listen to an audio clip of him on the Gridcast website.