Maybe I'm not exactly right in relation to NTFS file system, but I believe the basic idea is still accurate.
Found this on TIF:
With the latest release of Internet Explorer
<
and they're talking about IE4, but as we know IE7 is virtually the same>
and the Internet SDK,
it is expected that many different programs will take advantage of
the caching support provided. This means that there could be
thousands of URLs downloaded and kept in the cache.
Having thousands of URLs cached into a single directory causes
performance to degrade on computers running Windows 98, which uses
the FAT file system. There are two primary reasons:
- There is a linear search algorithm in FAT to look up a directory
entry.
- The in-memory structure of the directory entry caching used by
FAT does not work by directory but by individual file, causing
it to access the disk more often.
This is the main reason there are multiple directories in the
cache. Internet Explorer directs each URL into one of the multiple
directories (buckets), where the downloaded data is kept in the
form of a file. An index entry is created in the index file(s)
in that directory.
Temporary Internet Files
