sounds like lots of directions:

• why are duplicates such a frequent problem, sounds like upstream solutions are needed there?
• SSD would be faster read/write, yes (your data shouldn’t be on a single hard-drive, it should be regularly backed up at least - make the HDD the backup and copy the main database to SSD?); you might even consider a cloud service like AWS RDS
• for some use-cases, a noSQL database can be faster for reading - but it’s contextual

sounds like some changes would be a good idea 😅