IoT applications can be developed by domain experts using model-driven development environments. During the operation of such IoT applications, errors can occur that are not detected by existing monitoring systems, e.g., incorrectly measured sensor values. The goal of this project is to consider such error cases by means of data validation already during the model-driven application development and thus actively contribute to an improved error tolerance.
IoT environments show a variety of new error sources compared to traditional software applications. Causes include large device and hardware and software heterogeneity, high dynamics, due to moving, failing or newly added devices, or faulty or inaccurate sensors. In addition to software, hardware, and network errors that cause IoT applications to fail, operational errors can also occur. Here, operational errors are understood to be error cases that do not lead to the application crashing, but only to an undesired behavior, for example, due to an incorrect measurement of a temperature sensor.
Software, hardware and network errors can already be detected and handled by existing monitoring systems. Operational errors, however, are difficult to detect and handle, as undesired, perhaps even erroneous or even critical behavior comes to light despite error-free program execution. Especially in distributed IoT environments, the causes of operational errors are difficult to localize, since the individual devices and programs do not cause any abnormalities in monitoring systems. The time-consuming troubleshooting process results in high costs. These should be reduced by the work of this project by already generating suggestions during application modeling to take operational errors into account and thus actively contribute to an improved fault tolerance.
The micro project ACTION is carried out in the context of the Software Campus since 01.01.2022.
To be published:
Del Gaudio, Daniel; Ariguib, Boshra; Bartenbach, Arne; Solakis, Georgios: A live context model for semantic reasoning in IoT applications. In: Proceedings of the 18th Workshop on Context and Activity Modeling and Recognition (COMOREA) at IEEE Percom, 2022