Navigo: Interest Forwarding by Geolocations in Vehicular Named Data Networking

Giulio Grassi, Davide Pesavento, Giovanni Pau, Lixia Zhang, Serge Fdida

Abstract

Vehicular Ad-hoc Networks (VANETs) face significant challenges due to high mobility and frequent link disruptions. This paper presents Navigo, a geographic forwarding mechanism for Vehicular Named Data Networking (NDN). Navigo shifts the networking paradigm from node-centric to data-centric by binding data names to geographic regions. By routing Interests along road topologies using specialized "GeoFaces," Navigo improves reliability and efficiency compared to traditional IP-based geographic routing protocols like GPSR.

Introduction

Vehicular communication environments are notoriously difficult. In typical urban settings, network links between moving vehicles are fleeting, often lasting less than 10 seconds. Traditional IP-based networking, which relies on maintaining end-to-end paths to specific destination addresses, struggles with this constant "topology churn." When a vehicle moves or a building obstructs a signal, the path breaks, forcing the protocol to incur high overhead to re-learn the network. Navigo addresses this by moving away from finding specific nodes and instead focusing on finding specific data within geographic areas.

The Navigo Approach: Data-Centric Geography

Navigo leverages Named Data Networking (NDN), where users request data by name rather than by location. The core of Navigo is a system that binds these names to geographic regions using the Military Grid Reference System (MGRS).

GeoFaces and Geographic Naming

In a standard NDN router, a "Face" is a physical or logical interface. Navigo introduces "Geographic Faces" (GeoFaces), where each GeoFace represents a geographic destination. The Forwarding Information Base (FIB) maps data name prefixes to these GeoFaces, allowing the system to route requests toward specific grid squares (e.g., a 200x200m area).

Mapping data names to geographic regions via GeoFaces.

Road-Topology Routing

Instead of calculating paths based on volatile vehicle-to-vehicle links, Navigo utilizes the static physical road map. It employs a specialized Dijkstra algorithm where road segments are assigned costs based on lane counts, assuming wider roads are more likely to have vehicles available to relay packets. This ensures that Interests are forwarded along viable physical paths regardless of individual node movement.

System Architecture

The Navigo architecture integrates the NDN daemon with a specialized Link Adaptation Layer. This layer is responsible for the actual transmission of packets toward the destination area associated with a GeoFace and directing incoming data back to the corresponding GeoFace based on the provider's position.

Navigo architecture integrating GeoFaces with the Link Adaptation Layer.

Evaluation and Key Results

The authors evaluated Navigo using ndnSIM and SUMO traffic traces in a 2.1 x 2.1 km area of Los Angeles. The simulation accounted for urban obstructions using the CORNER radio propagation model. Navigo was compared against GPSR (Greedy Perimeter Stateless Routing), a standard IP-based geographic protocol.

  • Reliability: Navigo consistently satisfied more Interests than GPSR, maintaining a ~10% lead in success rate even under high demand.
  • User Experience: In music streaming scenarios, Navigo significantly reduced buffer underflows compared to GPSR.
  • Efficiency: Navigo proved more efficient, requiring fewer than 20 packets to satisfy an Interest, whereas GPSR often required over 30 due to ARP, ICMP, and Hello packet overhead.
  • Caching Benefits: Due to NDN's native caching, Navigo satisfied up to 40% of requests from nearby "data mule" vehicles, significantly offloading infrastructure compared to GPSR's negligible caching performance.

Conclusion

Navigo demonstrates that by routing toward geographic areas rather than specific nodes, vehicular networks can achieve much higher stability. By treating road topology as the primary routing substrate and utilizing in-network caching, Navigo provides a robust framework for both Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication.

Summary updated

BibTeX
@inproceedings{grassi2015navigo,
  author = {Giulio Grassi and Davide Pesavento and Giovanni Pau and Lixia Zhang and Serge Fdida},
  title = {{Navigo: Interest Forwarding by Geolocations in Vehicular Named Data Networking}},
  booktitle = {Wowmom},
  year = {2015},
  doi = {10.1109/wowmom.2015.7158165},
}
Figure