How to build vehicular networks in the real world

Abstract

There are now 1 billion vehicles in the world waiting to be connected to the Internet. At the same time, vehicular communication technologies have matured to a point in which massive deployment is both possible and feasible. One option for deployment is to wait for car manufacturers to embed DSRC/WAVE interfaces inside their latest models. However, since only 9% of the world's fleet is new every year, this would result in a time span of up to 20 years until 90% of the vehicles are finally connected. Another option is to rely entirely on cellular communications, such as GPRS, EDGE, 3G and LTE. This cellular only approach is impractical due to the capital expenses required for telecom operators to meet the demands of the impending tsunami of mobile data (expected to grow 1800% until 2016). Clearly, there is need for a low-cost wireless networking solution that can be placed in any vehicle and offers reliable connectivity, improved quality of experience and higher safety for drivers and passengers. This solution, we will argue, is vehicular mesh networking. Drawing from five years of research and our experience with a large testbed with hundreds of taxis and buses, currently under deployment in Porto, Portugal, we will address how city-scale deployment can be achieved and what kind of connectivity, bandwidth, quality of service and application support can be provided as the density of Internet gateways and the density of connected vehicles grows towards widespread deployment. Some attention will also be given to relevant issues in system design such as channel modeling, mobility patterns, networking protocols and large-scale simulation with manageable complexity. Finally, we will show how a vehicular mesh network can be used as a highly dense urban scanner, producing real-time data on-the-move, which can be leveraged from the cloud to help manage future cities, protect our environment and improve our quality of life.