Distributed processes dominate the computational landscape, from cloud computing to mobile apps, and on-demand access to spatial data has become routine. Web mapping services allow users to efciently view large data sets by accessing small subsets on demand, and dynamic services are a part of familiar applications like route navigation, weather forecasting, and trafc monitoring. Most products based on remotely sensed data are published in static form, but on-demand processing offers some promising new possibilities, particularly related to problems of scale. In this chapter, we explore some of the potential of on-demand processing, using downscaling of forest canopy structure as an example.