Enabling Location-aware Operation in Decentralized IoT Communications

We present an efficient design to enable location-aware operation in decentralized IoT communications. Large-scale IoT systems rep- resent the backbone of a smart city functioning, allowing pervasive environmental sensing across devices and networks. However, ex- isting IoT communication systems are largely driven by data types and miss out on embracing data location, which is fundamental in environment sensing. To address this issue, we demonstrate it is possible to efficiently embed a notion of location within the Zenoh protocol. We make it possible to steer message routing based on both data type and location, yet without altering the existing rout- ing core and message forwarding, unlike most existing solutions. We also present three encoding techniques for location data, each of them representing a different trade-off between expressiveness and performance overhead. Our evaluation uses a virtualized environ- ment and real-world packet traces of heterogeneous networks. We show, for example, that our design decreases the average message latency by more than 50% when routing data also based on location, while increasing throughput, compared to two different baselines.