It's only one of several interesting comparisons in the slides, but I think we need more data points on the disk space to make use of the data. Does it grow linearly? probably, but is there a +x in there? Mongo has some big upfront disk allocations. Were those considered somehow?