Scaling a Public Transport Monitoring System to Internet of Things Infrastructures
Haralampos Gavriilidis, Adrian Michalke, Laura Mons, Steffen Zeuch, Volker Markl
Proceedings of the 23rd International Conference on Extending Database Technology (EDBT 2020) | March 2020


Applications for the Internet of Things (IoT) face several challenges when it comes to exploiting the underlying infrastructure for data management operations efficiently. IoT infrastructures consist of heterogeneous compute nodes and geographically distributed network topologies. Today’s IoT applications offload data management to cloud-based stream processing engines (SPEs). However, this offloading represents a severe bottleneck that might hinder upcoming large-scale IoT applications in the future. In our demonstration, we showcase this problem using a public transport application as a potential large-scale IoT application. Our application consists of an interactive map for monitoring public transport vehicles and current demand. We implement this application on top of NebulaStream (NES), a new data management system that is designed for the IoT. In contrast to common cloud-based SPEs, NES answers queries by unifying cloud, fog, and sensor nodes under one system. Thus, NES minimizes network traffic and avoids resource over-utilization by considering the physical network topology and available compute nodes. The goal of this demonstration is to reveal the shortcomings of current system designs for large-scale IoT applications. Furthermore, we showcase how NES addresses these shortcomings and thus enables future large-scale IoT applications.