Context
We have a built-in caching library with a file system. We currently have performance issues due to the large number of records (e.g. up to 100,000). Problem: we save all fs entries in one "cache directory". Very large directories work poorly.
We are considering distributing these entries into subdirectories - as git does, for example. 100 subdirectories with ~ 1000 entries each.
Question
I understand that smaller directories will help with file system access.
But it will “propagate to subdirectories” speed up the movement of all entries, for example. listing / reading of all 100,000 records? That is, when we initialize / heat the cache from the FS storage, we need to move all 100,000 entries (and delete old entries), it may take 10 + minutes.
Will "data dissemination" reduce this "transit time." In addition, this “workaround” can actually / delete obsolete entries (for example, older than N days). Will Data Dissemination Improve Deletion Time?
Optional Context -
NTFS - Windows Family OS (Server 2003, 2008)
-Java J2ee application.
We / would appreciate any schooling on file system scalability.
Thanks in advance.
will be
p.s. , , , .