Slicer Grid Computing

Items discussed

• Stephen Aylward (KitWare) presented BatchMake interface to Condor
• Neil Jones (BIRN) presented (briefly) GridWizard
• Group discussed different data storage technologies: MIDAS, BIRN Data Repository/XNAT, etc.

Where we are

Both BatchMake and GridWizard have (partially working) demos of launching multiple processes on remote machines. However, there's still a lack of clarity of exactly what it is that the Slicer community needs. To that end, it's worth figuring out the answers to a few questions:

• What is the main community we should target? Core 1, 2, or 3?
• What are two specific use cases of a Slicer user interacting with batch processing facilities through Slicer? Where does data come from, where does it need to end up?
• What are two specific use cases of a NAMIC user interacting with batch processing facilities, perhaps outside of Slicer?
• Stephen (Aylward) presented a soup-to-nuts backend for managing "experiments", data, and displaying metrics computed from images in graphical form.
  • Does this fit in well with the above use cases?
  • If it is a requirement, who supports this? Stephen pointed out that the code is all in PHP and not terribly portable to new machines; this will need to be RPMd if BIRN is responsible for it.
• What look-and-feel of modules in Slicer is required? Do we need special "grid"-type parameters in the execution model, or should we just adhere to "min parameter", "max parameter", "step" notion?
  • Can we handle configuration of either tool out of band from Slicer?
  • Can we handle monitoring out of band from Slicer?

Until now, we have been content with infrastructure development: developing tools to perform generic batch processing. But it's becoming clear that we can't really make much more progress (aside from just improving the tools but not deploying them) without concrete goals.