Nod: Lightweight Continuous Neighbor Discovery on Everyday Devices

Nod: Lightweight Continuous Neighbor Discovery on Everyday Devices

연구 분야: Software Development

학회: International Conference on Mobile and Ubiquitous Systems: Computing, Networking, and Services

Continuous neighbor discovery (CND) allows wirelessly enabled devices to build a continuously updated picture of nearby devices. Such situational awareness is useful in many circumstances. While feasible on commodity devices like smartphones and wearables devices, CND is often not enabled due to practical limitations and concerns. Therefore, protocol implementations for CND have been mostly relegated to purpose-built sensor network deployments. However, the emergence of lightweight IoT devices and the need to rapidly discover what capabilities are around has renewed interest in CND on commodity devices. An important principle of CND protocols is their ability to be parameterized to achieve probabilistic guarantees regarding timeliness and assuredness of discovery events. However, transitioning from controlled deployments to commodity devices raises issues in providing guarantees related to CND. We explore two such concerns: (1) packet loss due to communication competition and (2) the time required to transition between sending and scanning for beacons. We propose a new CND protocol, NOD, that accounts for the impacts of these practical issues. We validate NOD through extensive simulations framed by real-world deployment settings and a proof-of-concept implementation for Android. NOD considers real-world application environments, allowing applications to identify optimal settings for guaranteeing probabilistic discovery in practical scenarios.