Geofences on the Blockchain: Enabling Decentralized Location-Based Services

Abstract

A decentralized ride-or carsharing application is among the early proposals of what smart contracts on blockchains may enable in the future. To facilitate use cases in the field of location-based services (LBS), smart contracts need to receive trustworthy positioning information, and be able to process them. We propose an approach on how geofences can be defined in smart contracts, and how supplied positions can be evaluated on whether they are contained in the geofence or not. The approach relies on existing location encoding systems like Geohashes and S2 cells that can transform polygons into a grid of cells. These can be stored in a smart contract to represent a geofence. An oracle run by a mobile network provider can submit network-based positioning information to the contract, that compares it with the geofence. We evaluate the location encoding systems on their ability to model city geofences and mobile network cell position estimates and analyze the costs associated with storing and evaluating received oracle-positions in an Ethereum-based smart contract implementation. Our results show that S2 encodings perform better than Geohashes, that the one-time cost of geofence definition corresponds linearly with the number of grid cells used, and that the evaluation of oracle-submitted locations does not incur high costs.