Need to better understand the employed objective functions for the planner,
investigate if the requirements put on participating triple-stores are reasonable,
explore if Avalanche can be changed to a stateless model, and empirically
evaluate if the approach truly scales to a large number of hosts. +