Energy-Efficient Multi-fragment Markov Model Guided Online Model-Based Testing for MPSoC

Term greenware capitalizes the energy efficiency, performance of processes and durability of hardware. Online monitoring/testing is a measure to support these features in autonomous mission critical systems (AMCS). The amount of time needed for sampling AMCS service quality modes has direct impact to energy consumption. We propose an efficient online testing method of AMCS service modes where the Multi-Fragment Markov Models (MFMM) are used for specifying the system reliability and quality related behavior on high level of abstraction, and the more concrete state and timing constraints are specified explicitly using Uppaal Probabilistic Timed Automata (UPTA). To interrelate these models we demonstrate how the MFMM is mapped to UPTA. Second contribution is the test case selection mechanism for online identification of AMCS service modes by model-based conformance testing. The efficiency of active mode sampling is achieved by serializing the test cases for each sampling session using hypotheses provided by MFMM. The hypotheses are tested using UPTA for online conformance. The approach is illustrated with the Bonfire Multi-Processor System-on-Chip (MPSoC).