I have NTFS folders that can grow to store between 100,000 and 1,000,000 files, the upper limit discussed in this answer to NTFS performance .
My files have the following characteristics:
1) They have long file names (usually from 64 to 100 characters).
2) For many files, the leading part of the file names may be the same for the first 20-40 characters.
Do long file names significantly affect the performance of an NTFS folder index when searching for a file with its name, index fragmentation, or index growth?
NTFS folder indexes (reportedly) B-trees. I tested my software with 50,000 files, but I have a "bon voyage" test with a small outflow of the file system. Testing up to 1,000,000 will require a week to run my software non-stop.
I was thinking of writing a simulator, but before I do this, does anyone have experience with this in the real world?
source
share