http://www.perlmonks.org?node_id=612737


in reply to Re: Perl solution for storage of large number of small files
in thread Perl solution for storage of large number of small files

32 chars names, lookup is always by name, full reads no seeks inside.

Files can be everything from 5-500K of data...

Concurrent writes to the storage, but never on the same file (as everything is passed through an index, serializing events, see above reply)

Sounds like you have an interesting solution in mind...
(This thread is becoming a fun-discussion, not only solving my problem but thinking about data-mining in general)