DOD is committed to
building net-centric systems (collaboration grids) so that data and information
are available where and when needed. The creation of information from data is
where a significant throughput issue
lies. Data volumes are expanding at a
tremendous rate, as is the complexity of algorithms used to process that data.
Yet, timelines are shrinking. The expectation is that once data is collected,
information should be readily available globally.
How throughput plays in a collaboration
grid is an interesting topic since we have traditionally thought of throughput
as a "stand alone" capability. As timelines
shrink, computational capability for collaboration grids is going to be needed
so that, in fact, information can quickly be derived from large volumes of data
using highly complex algorithms. That is why I think cluster processing grids,
distributed processing grids, and other ways to provide high throughput need to
be somehow integrated with collaboration grids and data center
grids.
What are your thoughts on that?
Thanks,
Peter.
Peter G. Raeth, Ph.D.
Research Engineer, Specialist in
Adaptive Automation
Ball Aerospace & Technologies Corp.
Advanced
Geospatial Group
2875 Presidential Drive, Suite 180
Fairborn Ohio
45324-6269 USA
praeth@ball.com937-320-6043 VOICE,
937-429-1687 FAX