Once we've done a full run of the workload on 4.2 and 4.4, we should compare the performance data and verify whether or not we see the performance improvement that we're expecting in 4.4.
The path after this ticket is where it starts to get fuzzy as it depends on what we observe. This ticket just tracks comparing the data but we should make additional tickets if we wish to do a deep dive into any particular aspect of the data.