RAS architectural description

From EGEE-see WIki

The model was designed to automatically adjust certain variables (using heuristics), giving them sensible values within a reasonable range of input parameters, relieving the burden from the user, and allowing him to concentrate on variation of actual and crucial model parameters. It seemed as a rational and logical step to enable modelling and achieving representative results even for users who are not the experts in POV-Ray SDL, broadening the range of potential users. The service takes, as its input, a POV-Ray scene, and a scene initialization file. It submits its rendering to the Grid, split into several jobs (“sections”), as desired. Some parameters are taken directly from the POV-Ray scene initialization file, while some are specified on the command-line as job-submission parameters. The section jobs can notify the user when they are done via e-mail. After all processes are done, the rendered frames are collected, and can be joined together in Animated GIF or AVI/Xvid format. The user is protected from Grid internals, which are wrapped inside the application scripts.

RAS workflow Specific requirements:

• POV-Ray (Persistence of Vision Ray-Tracer)
• mencoder
• imagemagick