POLICY :: Next Generation Internet > Grids (archived)

This page was archived when this portal was restructured in early 2007. For up to date information, please visit the following sites:

Pervasive and Trusted Network and Service Infrastructures, the first Challenge under ICT research in the Seventh Research Framework Programme (FP7, 2007-2013)
IST Research under the Sixth Research Framework Programme (FP6, 2002-2006):
- Grids for Complex Problems Solving
- eInfrastructures: Building on GÉANT and Grids

Building Grids for Europe

Building a commanding position in grid computing is critical for Europe. By allowing anyone, anywhere to easily access supercomputer-level processing power and knowledge resources, Grids will underpin progress in European engineering and business. Their impact on European society may be just as profound.

Grid-based technologies today hold the potential to revolutionise the Internet of tomorrow as the WWW did for the Internet of today

Everyone understands that electricity is generated using everything from nuclear plants to windmills. We also know it is stored, transmitted, distributed, sold and re-sold in complex commercial and technical arrangements.

But we don't need to know any of these details - electricity is just something that comes out of a plug. It's available, affordable, and dependable. Every now and then we pay a bill, but we control how much or how little we use, and can choose the best supplier for our needs. Most of us don't even know what electricity really is - it's just there, making things work.

In the knowledge-based economy that Europe must become if it is to prosper in the 21st century, we need a second grid: one that supplies enormous computing power and knowledge without forcing us to worry about where and how.

What is a Grid?

This new sort of grid works by connecting together huge varieties of computers, data repositories, software programs, scientific instruments and more. Connecting equipment together, in itself, is not revolutionary - after all, every time you look up a page on the Web you are connecting your PC to a computer somewhere else in the world.

A grid, however, allows you to tap that computer's processing power, rather than simply look at some of its files. Now add the power of a thousand other computers spread across the globe. You choose how much power you want to use - your Computational Grid parcels the job up and shares it out to all the computers connected to it. Suddenly your PC has become a gateway to a supercomputer, but you only pay for the power you need, when you need it.

But you have much more than that - you can also access massive amounts of data, whether it is stored in dedicated data storage devices or flowing from scientific instruments and sensors. They are all attached to your Data Grid which - thanks to the computing power available - can process as well as store this data easily, using a whole range of software designed to make the most of the Grid's computing power.

And more still - Grid computing power enables new sorts of software to become possible. Instead of data flowing coming out of that plug in the wall you now get knowledge, created by semantic analysis software running on your Knowledge Grid. Your PC can understand and manipulate this knowledge, providing you with instant access to answers from across the world.

Finally, you are not alone on this Grid - through it you work with your project team, composed of members located in different continents, working for different organisations. This Service Grid creates a virtual organisation, working in a virtual laboratory. Sensors and experiments are controlled remotely, data is stored, shared and processed instantly, results are analysed on-line, and relevant knowledge from the world's databases found and combined with it.

It doesn't matter if your team is modelling the Earth's atmosphere, designing cars, creating animated films or finding new medicines, the basic principle is the same: your grid supplies all the computing power, software, data and knowledge you need in one integrated package, and helps project teams work more closely together.

From Laboratory to Industry

EU-funded research projects develop Grids which have already been 'stress tested' in a demanding, multi-location environment.

Grids were therefore originally built for 'big science' projects, involving large amounts of data and highly intensive calculations. The Grid was in fact pioneered in Europe's CERN particle physics laboratory, home of the World Wide Web. While Grids are now being built all over the world, they are uniquely important to European science, as they will allow the continent to better pool its scientific resources to an unprecedented degree.

Today, however, Grids are entering completely new waters. Their impact on the world of engineering and business will be massive and is already being felt - but sectors as diverse as medicine, culture and entertainment are also facing radical changes.

By making supercomputing affordable to organisations who only need this sort of computing power occasionally, for example, Grids will empower Europe's small and medium sized enterprises (SMEs). And by allowing users to share data, software, instruments and generally work together more closely across greater distances, they will have a significant impact on Europe?s manufacturing industries, where teams in different companies work together on highly intensive design and engineering projects.

Technical Challenges

Today's Grids, however, are complex to create, requiring expert resources in operating systems and wide-area networking to deploy. They are also difficult to program, as they span numerous platforms and protocols, and cumbersome to maintain - some need up to 20 personnel to restart after a crash.

The vision of the Next Generation Grid is that it should eventually "disappear", becoming as pervasive, robust and casually familiar as electricity or the telephone service. This means meeting a number of technical challenges:

Poor reliability is not acceptable to paying business users, and is certainly not good enough for critical systems in areas ranging from finance to nuclear power stations;
"Hiding the Grid": While the research world is staffed with world-class computer scientists who intimately understand what they themselves have built, this will never transfer to the wider world. The Grid needs to be 'virtualised', the complex resource brokering and management going on behind the scenes hidden behind a user friendly layer of standard, interoperable interfaces;
The research world also worries a lot less about security - grids were originally developed to help researchers share data and solve problems together. Data and resource sharing takes place in industry, but it needs to be controlled through access, trust and authentication systems;
User-friendliness: Virtualisation won't be able to cover every conceivable usage of the Grid. Users will regularly need to define their own requests. Easy-to-program interfaces, such as the "Wizards" used in desktop computing, will be required;
Standards: Finally, most Grids today are built around an early set of de facto US standards and toolkits. A new generation of global standards are therefore required to fully exploit the Grid?s potential.

A European Challenge

Grids are therefore a critical enabling technology for scientific excellence and industrial competitiveness. Their importance was underlined in the EC's current Sixth Research Framework Programme (see the Research & Innovation theme), which doubled the budget available for Grids research and focused Europe's efforts in two main areas:

building grid technologies for solving complex problems ranging from industrial design to environmental modeling - see Grid-based systems for Complex Problems Solving for an overview of Grid activities in the current and previous Research Framework Programmes;
e-Infrastructures: powerful research infrastructures to help Europe pool its scientific and technical resources: see the Research Infrastructures site, particularly eInfrastructures: Building on G?NT and Grids

The EU drive to promote the adoption of IPv6 is also critical, as Grids need this more advanced internet protocol if they are to fulfill their promise.

The European Commission, of course, is not alone in this view, with several EU Member States launching national Grid research programmes. By actively ensuring better coordination of these national and European research activities (see the workshop of July 2003), the EU aims to create a ?European Research Area? in the Grid domain and build Europe's global leadership in this critical research area.

Updated 11 March, 2004.
DG Information Society