Hi everyone!
What is the best way to split massive amount of files (tens of millions)
among folders? What would be the best files-per-folder ratio, hierarchy
depth and other params in order to get better random file open
performance and yet not waste too much disk space with folders? Has
anyone ran this type of analysis?
Another question: how justified would be to have a filepath->fileID map
in order to have most often used files being opened via ID? Is there
going to be any significant performance improvement?
All these Qs are primarily related to NTFS.
TIA,
Vladimir