Collect data as the network scales up beyond 50 nodes (what we currently run at). And as we increase the size of blocks. For each release, we want to produce a "table" of performance.