Calculating worst-case execution times of transactions in databases for event-driven, hard real-time embedded systems

In embedded real-time systems the use of databases increases, particularly to ensure the quality of data and to simplify the development of software. Especially for safety-critical applications one must guarantee in advance by suitable verification-algorithms that all deadlines hold in all possible cases of use. The worst-case execution time of transactions is an essential input of those algorithms. Therefore, we present here a method to calculate these times for non-distributed, main memory databases in event-driven, embedded hard real-time systems. We use the very powerful means of event-streams and event-dependency-matrices for a detailed description of the embedding environment and an application model to specify the necessary information about the application itself.