Embedded Database Management Performance

Current methods for developing software for mobile devices don´t account for input/output (I/O) options when design decisions are made. I/O needs are typically addressed in an ad hoc manner, and inexperienced application developers are not accustomed to programming for resource constrained devices and tend to ignore differences in performance between various persistent storage options. In this paper, we investigate the two primary methods for persistent storage on mobile devices and use a battery of tests to evaluate the optimal ways in which to retrieve persistent data from mobile devices. In our investigation, we uncover that if the number of records are less than 1000 then sequential files should be used as the data storage option. Furthermore, in cases involving nonindexed columns, the clear winner appears to be the embedded DBMS solution, especially as the data scales.