Performance assessment of vehicular delay-tolerant networks

Abstract

Vehicular networks have attracted considerable attention in the last few years, both in academia and industry. One of the main reasons for their growing popularity can be attributed to the various applications that they make possible. Road safety, traffic monitoring, driving assistance, entertainment, and delivering connectivity to rural/remote communities or catastrophe-hit areas are just a few examples of the many applications envisioned for these networks. Nevertheless, the special characteristics of vehicular networks such as high mobility, highly dynamic network topology, short contact durations, disruption, intermittent connectivity, significant loss rates, variable node density, and network partitioning introduce unique challenges, which greatly impact the deployment of these networks. Such challenges make data dissemination and routing interesting research topics within the vehicular networking area, which are addressed by this research. The work presented in this thesis is motivated by the need to find new solutions to the communication problems arising in disconnected opportunistic vehicular networking scenarios. A new network architecture for vehicular communications is therefore proposed in this work, called vehicular delay-tolerant network (VDTN). Its layered structure is introduced and the corresponding performance analysis is conducted. This architecture differs from other proposals in the literature in several respects. Briefly, it adopts a store-carry-and-forward paradigm combined with an IP over VDTN approach and out-of-band signaling with control and data plane separation. This thesis presents studies on the impact of stationary relay nodes, node density, vehicles movement patterns, and storage constraints on the VDTN network performance, in terms of bundle delivery probability and bundle average delivery delay. Of particular interest to this thesis is the performance improvement of these networks. In particular, node localization information is exploited to improve and optimize the use of data plane resources. It is demonstrated the performance gains attainable in a VDTN through the cooperation between network nodes, in terms of bundle delivery probability and bundle average delivery delay. These performance metrics are also used to investigate the impact of non-priority and priority-based queueing disciplines. Finally, a detailed analysis is performed with the proposed routing protocol for VDTNs, called GeoSpray, against popular single-copy (Direct Delivery, First Contact, GeOpps), and multiplecopy (Epidemic, Spray and Wait, PRoPHET) routing protocols. Such protocols are considered reference in the literature of DTN networks and were deployed in VDTNs. It is shown that GeoSpray yields significant performance gains in terms of the bundle delivery probability and the bundle average delivery delay. The proposed protocol proves to be efficient in terms of storage and bandwidth resources utilization. The results presented in this thesis are based on computer simulations and testbed experiments. The lack of a simulator specialized for VDTN layered network architecture, created the necessity to propose, develop, and implement a simulation tool for VDTNs, called VDTNsim. VDTNsim was used as a basis for the creation of a prototype of a VDTN laboratory testbed, named VDTN@Lab. This thesis aims to contribute to the advance of the state-of-the-art on techniques for tackling the challenges that arise from the unique properties of vehicular networks. Further, this thesis highlights important guidelines for the improvement and design of new protocols, algorithms, services, and applications for vehicular delay-tolerant networks.

O presente resumo alargado em língua portuguesa sintetiza a tese de Doutoramento intitulada “Avaliação do Desempenho de Redes Veiculares com Ligações Intermitentes” (Performance Assessment of Vehicular Delay-Tolerant Networks). Começa-se por enumerar os objectivos e referir as principais contribuições desta investigação. Depois são sintetizadas as principais conclusões e apontadas direcções para trabalho futuro.