Montage is an application developed as part of the National Virtual Observatory (NVO), a major astronomy project. The overall goal of the NVO is to enable new science by greatly enhancing access to data and computing resources. NVO makes it easy to locate, retrieve, and analyze data from archives and catalogs worldwide. Within this framework, the Montage project provides a portable, compute-intensive service that delivers science-grade custom mosaics on demand, with requests made through existing portals. Science-grade in this context requires that terrestrial and instrumental features be removed from images in a way that can be described quantitatively; custom refers to user-specified parameters of projection, coordinates, size, rotation, and spatial sampling.  Three sample (not science-grade) images are shown below; the gray lines illustrate the edges of the images that have been brought into the mosaic.

This image is an full resolution (1.5 arcminute pixels) mosaic of the 60 micron IRAS image (ISSA) data. It covers a 60 degree square region of the sky centered on the Andromeda Galaxy (Messier 031).	The plane of the galaxy runs across the top of the image and the active area below that coincides with the visible constellation of Orion (his head is in the general vicinity of the ring of brightish emission on the right and his belt runs through the very bright area on the left). The mosaic is comprised of 31 original ISSA plates covering an area 45 degrees across.	Rho Ophiuchus, 45 degrees wide in Galactic coordinates, captured in 32 images. The Galactic center is bottom center of the image.

All images courtesy of Montage project.

From the VGrADS perspective, computing a mosaic is a workflow computation. The figure below illustrates the steps in this workflow. In a realistic example, the various levels would be much wider (typically as wide as the number of original images in the mosaic) and possibly have differing widths. Also, the computational requirements of the individual nodes might differ, depending on the amount of adjustment required.



In summary, Montage represents a more complex set of workflows than EMAN. It has provided valuable insights into the requirements that such workflows place on grid schedulers.