Tuesday, April 9, 2013

The European Grid Infrastructure Community Forum kicks off in Manchester

The European Grid Infrastructure Community Forum kicked off in Manchester this morning. Manchester is famous for many things (architecture, culture, music, science and engineering to name just a few) but arguably is recognised most widely for a certain football team or two. There is a distinct football theme to the artwork for this week’s event, and even Professor Nancy Rothwell, Vice Chancellor of the University of Manchester, recommended in her welcome address that we refrain from discussing it with locals if at all possible.

Peter Coveney, Director of the Centre for Computational Science at University College London presented recent developments in e-Infrastructures that are helping to unlock new science. He outlined the challenges he sees in putting the rapidly growing body of scientific data to use and particularly focused on applications in biomedicine and materials science research. Discussions were around how to fund these activities and Coveney gave some examples of recent UK funding calls.

Kostas Glinos, Head of the e-Infrastructures Unit at the European Commission focused on Horizon2020, which is the European level funding programme set to start in 2014. Glinos noted that the approach to funding will be integrated and service driven. The aim is to provide integrated, de-siloed services and to establish synergies with individual member state’s e-infrastructure development programmes. Horizon2020 will move beyond science to include the science-industry-society triangle, mapping out societal and policy needs and integrating innovation into the process. Key to success is financial sustainability with ‘proper’ positioning on the user pays / provider pays axis, and a governance structure that maximises EU added value. The Commission has several priorities in the e-Infrastructure arena (to stick with the football theme). They are interested in big data, with seamless and open access for researchers. Their strategy for computational infrastructure covers supercomputing and open computing, such as grids and clouds. The funding will also focus on implementing the recommendations of the GÉANT expert group.

As a result, the first projects in Horizon2020 should produce data centric science and engineering, community driven data infrastructures and global coordination for research data. The programme will also aim to fund a series of centres of excellence that will provide training and skills development for HPC. Essentially, the three pillars of their HPC strategy include next generation exascale computing, provision of the best facilities and services for academia and industry, and should achieve excellence in HPC applications. They are also considering Public Private Partnerships to link demand and supply. According to the discussions though, the European Commission should make sure to support researchers to find ways of working with the data. In between gathering, storing and ensuring the long term archiving and labelling of data in an accessible way, researchers actually need to be able to handle the data successfully to produce results in the first place.

Steven Newhouse, Director of EGI, closed the session with a look at the road ahead. “Where have we come from and where are we going?” asked Newhouse. He looked back at the journey from the web to the grid to the discovery of the Higgs particle and asked, where do we go from here? Back in the days of the European Data Grid ten years ago, the grid boasted 4000 cores and 30 sites. In 2013, the infrastructure includes 400,000 cores and 350 resource centres. For EGI, the three areas of innovation that will take the world's largest production grid infrastructure into the future are in community and coordination, operational infrastructure and Virtual research environments. With one year to go before the final whistle blows on the EGI-InSPIRE project, but at least 18 months until Horzon2020 projects start warming up, there are some tough questions to be asked – and answered. The focus has to be on the critical core services - who needs what, and how can we sustain what we currently have? Discussions have been underway in the community for some time, and no doubt will continue intensively this week.

One thing is clear however, developing our human network is vital. You can’t win the game without the right players on the pitch. EGI already fields a network of National Grid Infrastructure Liaisons who focus on supporting the user community from inside the EGI-InSPIRE project. This Community Forum is the first EGI conference that also welcomes EGI’s new champions, a hand-picked team of researchers hailing from communities such as engineering, biomedical science, astronomy and beyond, who are supported financially by EGI to promote grid computing and EGI within their community. Several of them are at the event and you can find them online at If you bump into any of them, please say hello to our latest players and welcome them to the field!

No comments: