Feasibility study: cross-back-end process graph handling

[soxofaan]

from #52:

Feasibility study of cross-back-end processing: Support process graphs with collections and processes from different backends

Increasing levels of complexity to evaluate:

• basic POC: find maximal subgraphs supported by each backend (based on collection and/or process availability)
• split a given process graph into valid subgraphs
• manually run a workflow that combines sub-process-graphs run on separate back-ends
• ultimately (when deemed feasible): implement automatic execution through "partitioned job" system in aggregator: split, schedule (with dependencies), combine, ...

[jdries]

The feasibility study was performed and presented at SRR4. I'll close this issue now, as it's replaced with more specific issues such as setting up an actual POC.