Real-time driver monitoring systems (DMS) are increasingly mandated in modern vehicles to improve road safety by detecting driver distraction and fatigue. These systems generate continuous data that must be processed, stored and queried with minimal latency on resource-constrained embedded platforms. This study presents an experimental evaluation of three widely used database management systems (DBMS) PostgreSQL, MongoDB and Redis for handling driver-monitoring data on a Raspberry Pi 4 to emulate in-vehicle embedded systems. A SmartEye camera produces driver monitoring data at 60 records/s, which are inserted into each DBMS while simultaneously performing reads and deletes. Redis was fastest on all operations. PostgreSQL was the slowest for all operations, with MongoDB between the two DBMS.