Both of these questions seem to have an implicit assumption that we are talking about mapping applications to "idle desktops" or some similar environment with poor reliability and I/O capability. As Chris points out, grid technologies can also be used to manage pools of highly reliable and well-connected servers. Perhaps as well as answering the questions as below, we should address the misunderstanding that seems to be inherent in the question. For (a), as a further example to that specified by Chris, we could be dynamically allocating CPUs used to host virtual desktops. With respect to (c), Ioan Raicu is doing some nice work in which he greatly *increases* I/O capability using grid techniques, by dynamically allocating an increasing number of servers to a data analysis workload as the amount of work increases. Chris Kantarjiev wrote:
3. Applications a. Which non-compute intensive applications are well suited to running on a grid?
There's a wide range of transactional, three-tier applications that are response-time sensitive, that are well suited to running on a grid - where the grid infrastructure's flexibility and dynamic reconfigurability can be brought to bear.
For example, a "web store" can use grid technologies to repurpose the servers in a pool to respond to ebbs and flows in customer demand, both at the HTTP-serving leaves and the middleware engine layers.
Grid-enabled database servers and storage units can similarly respond to increased demand to spread the load of additional customers or quarter-end tasks. When the load subsides, the storage can be deprovisioned and put to use elsewhere.
c. Are I/O intensive applications not suitable for grid?
With a suitable storage infrastructure, I/O intensive applications can be supported by grid-enabled servers and storage structures that spread the load as needed to provide the promised or requested response time ... at a price, of course!
_______________________________________________ tsc mailing list tsc@ogf.org http://www.ogf.org/mailman/listinfo/tsc
-- Ian Foster, Director, Computation Institute Argonne National Laboratory & University of Chicago Argonne: MCS/221, 9700 S. Cass Ave, Argonne, IL 60439 Chicago: Rm 405, 5640 S. Ellis Ave, Chicago, IL 60637 Tel: +1 630 252 4619. Web: www.ci.uchicago.edu. Globus Alliance: www.globus.org.