Elsevier

Vehicular Communications

Volume 5, July 2016, Pages 35-43
Vehicular Communications

Characterization of a delay and disruption tolerant network in the Amazon basin

https://doi.org/10.1016/j.vehcom.2016.09.002Get rights and content

Abstract

Most cities in the Amazon have no data communication infrastructure, and rivers are most of the time the only access mode to connect small cities to urban centers. In this paper, we investigate the deployment of vehicular ad hoc networks (VANETs) formed by boats along the rivers of the Amazon as an alternative to interconnect the small cities to the capital Manaus. Given that boats will opportunistically connect to each other, we have the scenario of a delay and disruption tolerant network (DTN). In a nutshell, the idea is to take advantage of the contacts between the boats that travel daily transporting passengers and products to small cities, communities, villages, as well as to other states. We evaluate the capacity of a DTN formed by boats, using simulations with real boat traffic and mobility traces captured in the Amazon Basin as input. Results show that potential of raw data transfer in this large-scale DTN can achieve 1.38 TB over a week for some 45 hours of contact time between boats.

Introduction

The Brazilian Amazon has the largest river basin in the world, with 20% of the planet's freshwater, comprising over 20,000 km of navigable waterways. The Amazon represents over 40% of the Brazilian territory, which corresponds to more than five times the size of France. However, the Amazon region faces several problems of scarce data communications infrastructure, which affect the level of socio-economic development of the region. The Amazon's geographical characteristics favor the river transportation [1], which is practically the only mode for passenger and cargo transport in the region. Airplane travels between small cities are used, often with multiple stops, increasing operational costs. One major disadvantage of boat transportation is the travel time. Fluvial distances are frequently given in days. For example, 3.5 days from Manaus to Belém and 5 days the other way around (the difference being due to the boat going down or up river).

One alternative to cope with the lack of telecommunications infrastructure in the Amazonian cities is the use of boats in communications systems, in a similar way as VANETs [2], [3], [4], [5], [6] use cars [7], [8], buses [9], [10], and even drones [11] to build the communication system. A large number of cities, communities, villages and states would benefit from VANETs formed by boats that regularly travel in the Amazon basin. Considering that there are over 45,000 of such boats, with a pre-existing infrastructure of ports and terminals, boats can use opportunistic contacts to send messages from one city to another and to exchange data between each other along the river bed, increasing the capillarity of this network. Nevertheless, there are challenges that add to the day-long boat travel times, such as, wireless communication behaves differently in indoor or urban environments, characterizing contacts between boats as well as their trajectories in the river beds. Therefore, this paper investigates a data communication network for the Amazon cities currently living with precarious or no data communication infrastructure. We analyze a VANET implementation in a real scenario of boats navigating in the rivers of the Amazon. Thus, our main goals are: (i) to characterize data transmission of boats in the Amazon, (ii) to characterize the contact times between boats, and (iii) to calculate, based on real data, an estimate of the capacity of data transfer through simulation. In summary, the main contributions of this work are:

  • new mobility traces from boats in the Negro and Amazonas rivers.

  • the estimation of DTN [12], [13], [14], [15] capacity along the Amazonas and Negro rivers, through computer simulations. Results show that the potential of data transfer between boats is in the order of terabytes over one week.

This paper builds on our previous work [16], where we conduct experimental measurements and derive preliminary estimations of the DTN capacity, based on the assumption of Uniform Rectilinear Motion of boats. In contrast, in this work we use real mobility traces obtained in Baixo Amazonas river channel. Then, by simulating the movement of more than a hundred boats in this channel, we measure the capacity of this DTN. The simulation parameters are tuned according to the values measured in the practical experiments.

This paper is organized as follows. Section 2 presents related work, while Section 3 describes the methodology used in this work, the practical experiments between crossing boats in the Negro river, and the boat mobility trace captured with GPS in a real scenario. The information obtained from the experiments is then used in the simulations presented in Section 4. Finally, Section 5 concludes this manuscript and discusses future work.

Section snippets

Related work

Several papers evaluate the capacity [7], [16], [17], [18], contact times [19], inter-contact [20], [21], [22] times and the mobility [23], [24], [25] of VANETs and DTNs. Nevertheless, few papers analyze the data transfer capacity of mobile devices in river scenarios.

Zhang et al. [26] investigate the bus-to-bus contact times and its impact on DTN routing performance in the city of Massachusetts, United States. Buses follow well-defined routes and find other buses that establish pair-wise

Methodology

Initially, we collected data about contact time and amount of data transferred, in experiments in the Negro river with crossing boats. These experimental values are used to define different simulation parameters. Then, we captured real traces of boats navigating in the region, in order to represent the mobility of boats in these networks, similarly to [19] or [35]. Due to time and financial constraints, we limited the scope of our measurements to the Baixo Amazonas channel, shown in Fig. 2. We

Simulation of DTN in the Amazon basin

Considering the boat navigation scenario in the Baixo Amazonas channel, the objective of this work is to characterize boat contact times and to evaluate data transfer capacity of these contacts through simulation. The simulated network has a set B of n boats, i.e., B={b1,b2,b3,...,bn}. Our example scenario of the Baixo Amazonas has n=109 boats. The contacts are characterized by boat pairs, i.e. (bi,bj) where ij, bi,bjB. Boat pairs (bi,bj)=(bj,bi), are taken into account only once in the

Conclusions and future work

The Amazon, in the north of Brazil, corresponds to almost 40% of the Brazilian territory. In this area, water transport is predominant since most cities and villages are accessible only by boat. Because of the difficulty of access, the communication infrastructure is scarce in most of the region. In such context, VANETs based on boats appear as an interesting alternative for providing communication capabilities to the Amazon region. Therefore, this work investigates the potential of this

Acknowledgements

This work was partially funded by CAPES, CNPq, Fapeam, Faperj, Muraki Foundation, and GE Global Research – Brazil. The authors also thank the contribution of Dr. Marcelo Dias de Amorim, Dr. Miguel E.M. Campista, Eliézer P. de Moura, government agencies SEMED-AM and CBMAM who kindly provided the boats used in the experiments. We would like to thank support of boats that have been made real traces, government agency AHIMOC that given us scheduling of boats of waterways of Western Amazon, and

References (39)

  • J. Burgess et al.

    Maxprop: routing for vehicle-based disruption-tolerant networks

  • K. Fall

    A delay-tolerant network architecture for challenged internets

  • M. Demmer et al.

    Implementing delay tolerant networking

  • M. Demmer et al.

    Delay tolerant routing for developing regions

  • A.V. Vasilakos et al.

    Delay Tolerant Networks: Protocols and Applications

    (2011)
  • A.J. Santos et al.

    Capacity analysis of a delay and disruption tolerant network in the Amazon basin

  • V. Soares et al.

    Evaluating the impact of storage capacity constraints on vehicular delay-tolerant networks

  • Y. Huang et al.

    Multicast capacity analysis for social-proximity urban bus-assisted VANETs

  • G. Marfia et al.

    VANET: on mobility scenarios and urban infrastructure. A case study

  • Cited by (0)

    View full text