Abstract
The emergence of edge stream processing has created a new way of processing real-time data from the Internet of Things (IoT), which comprises a plethora of geographically dispersed physical devices equipped with sensors and actuators that exchange data with the Cloud. Nevertheless, edge stream processing systems face new challenges, including dynamic workloads, resource limitations, and multi-tenant application hosting. Adaptive resource management has been proposed to address these issues. However, this technique may lead to Service Level Objective (SLO) violations when the system encounters resource constraints. To mitigate this problem, we investigate the benefits of using priority-based stream data to reduce the SLO violations associated with adaptive resource management. Our findings demonstrate that segregating data according to their priority levels and processing them accordingly can significantly enhance the efficiency and stability of the system. We implemented this technique on the Storm Streaming system and used RIOT as a benchmark, employing VRebalance and other approaches to adjust system resources dynamically.