Computer-centric problems are the domain of High performance computing. The user needs "teraflops", as many as possible. Many computer-centric applications can benefit from the Grid to combine large computational resources in order to tackle problems that cannot be solved on a single system, or at least to do so much more quickly.

Data-centric problems - also called data-intensive problems - are the primary driving force behind the Grid at present, and will likely continue to be for some time in the future. Over the next decade, huge amounts of scientific data will come from everywhere: not only particle and astrophysics experiments, but also sensors monitoring just about everything you can imagine - from precise maps of motion of the earth's crust to highly localized weather data. The Grid will be used to collect, store and analyze data maintained in geographically distributed repositories, digital libraries, and databases. Some examples include:

• Future high-energy physics experiments will generate terabytes of data per day (link to Grid@CERN). Thousands of physicists from hundreds of research institutes, laboratories and universities worldwide will need to access and analyze them.
• The Digital Sky Survey will make many terabytes of astronomical photographic data available in numerous network-accessible databases. This facility will enable new approaches to astronomical research based on distributed analysis [link to examples e.g. recent Economist article…].
• Modern meteorological forecasting systems make extensive use of data assimilation to incorporate remote satellite observations. This involves the movement and processing of many gigabytes of data.

In contrast, community-centric problems, also referred to as collaborative applications, are concerned primarily with enabling and enhancing human-to-human interactions, attempting to bring people or communities together for collaborations of various types. They are often structured in terms of a "virtual shared space" which enables the shared use of computational resources such as data archives and simulations (of course, they also have characteristics of the other application classes just described).

Examples range from interactive video presentation and conferencing from many sites simultaneously, to distributed musical concerts, to supporting collaborations of dozens scientists, engineers, and mathematicians around the world trying to perform together a complex simulations as data flow from detectors of various kinds.

Some of these applications require, or are enhanced by, real-time user interaction of many forms, from decision-making to visualization. Such real-time requirements, imposed by human perceptual capabilities and the rich variety of interactions that can take place, are some of the challenging aspects of collaborative applications from a Grid perspective.

The requirements for responsiveness are often in direct contradiction to the needs of highly distributed computer systems. However, the Grid can provide some benefits to these applications, when they need underlying computer power or data storage to facilitate the shared experience. Here are some examples:

• The BoilerMaker system developed at Argonne National Laboratory allows multiple users to collaborate on the design of emission control systems in industrial incinerators. The different users interact with each other and with a simulation of the incinerator. The simulation itself can be updated more quickly using the grid technology.
• The NICE system developed at the University of Illinois at Chicago allows children to participate in the creation and maintenance of realistic virtual worlds, for entertainment and education. Again, by distributing these virtual world simulations on a Grid, more users can benefit from the system.