Estimations of execution time are essential for design and development of safety critical embedded real-time systems, such as avionics, automotive and aerospace systems. In such systems, execution time is part of the functional specification, hence correct behaviour requires sufficiently powerful target hardware to meet deadlines or achieve required polling rates, etc. Yet, grossly overestimated resource usage results in excessive cost per unit. For a proper choice of the target platform, qualitatively good execution time estimates are required at an early stage of the development process. In this paper we propose a framework which provides software engineers with execution time estimates of the software under development in a demand-driven manner, i. e., the engineers ask for timing information at program or function level with respect to different target hardware platforms. In a platform-independent manner we extract the necessary information from the code and combine it with platform-specific information, resulting in the time estimate. We implemented our framework on top of the test input generator FSHELL and its query language FQL. Preliminary experiments on C code show the viability of our approach.