Towards modular average-case timing in real-time languages: an application to real-time Java

In certain real-time systems worst-case execution time estimates often lead to a waste of resources. In hard real-time systems these types of estimates are essential to guarantee temporal requirements are met. However in soft-real time systems using other measurements, such as average-case timing, to complement the worst-case estimates can lead to better utilisation of resources while ensuring most, if not all, deadlines are met. We propose a methodology to integrate modular average-case timing in Real-Time languages. Previously statically determining average-case time, if possible, required rigorous mathematical techniques. Our approach, which is based on a new programming paradigm called MOQA and is built on Real-Time Java, simplifies the process and allows for automation.