YetiOS: an Adaptive Operating System for Wireless Sensor Networks

Recent decades have experienced a breakthrough in the study of Wireless Sensor Networks (WSNs) as a key technology for Internet of Things (IoT). Operating Systems (OSes) for resource-constraint WSN devices have been developed during the last years, facilitating application development for these networks. However, these OSes are static software modules and do not take into account environmental changes to optimize themselves. In this paper we present YetiOS, a novel OS that can change its properties dynamically through an adaptive core. This way, the OS adapts itself to the environment and can optimize the performance, power consumption, security or tasks latency on different scenarios while maintaining the same developing interface. YetiOS also includes a dynamic layered network stack architecture, allowing different sensor network protocols to be implemented. Besides, it features Linux-like device drivers, a FreeRTOS based process management module, an I/O interface module and an energy management module. Performance tests presented in this work demonstrate YetiOS suitability for low-power resource-constraint WSN nodes. In addition, a use case of the adaptive core is presented, and its performance test demonstrates the dynamic optimization achieved.