Optimizing the executiontime o fSQLite on an ABB robot controller

Abstract: The purpose of this thesis is to generate a set of general guidelines to improve performance of the SQLite database management system for use as a component in the main computer of ABB industrial robots. This was accomplished by measuring the factors that affect query execution time and comparing SQLite features to systems it could potentially replace: chiefly ABB’s configuration storage system and real-time logging system. Results show that SQlite is slower than ABBs current configuration system and real-time logger. A series of tests were carried out that give a rough description of SQLites performance. Read and update become significantly faster when used with an index, write becomes a bit slower. In an indexed database the number of rows is not important; in a database without an index they cause significant slowing. Surprisingly, increasing the number of threads concurrently accessing a database had only a mild effect on timing.