No full text
Doctoral thesis (Dissertations and theses)
Routing Strategies and Content Dissemination Techniques for Software-Defined Vehicular Networks
di Maio, Antonio
2020
 

Files


Full Text
No document available.

Send to



Details



Keywords :
SDN; SDVN; Software-Defined; Routing; Content Dissemination; Congestion; Throughput; Fairness; Delay; Graph; Multi-channel; Scheduling; Performance Evaluation; MANET; VANET; DSDV; GPSR; AODV; Unicast; Multicast; Connected; Vehicular; Networking; Networks; V2V; Centralized; Controller; Load-balancing; IEEE 802.11p; IEEE 1609.4; DSRC; WAVE; Heuristics; Simulation; OMNeT++; SUMO; INET; Veins; Cellular; 5G; Manhattan; Urban; Scenario; LuST; Luxembourg; Mobility
Abstract :
[en] Over the past years, vehicular networking has enabled a wide range of new applications that improve vehicular safety, efficiency, comfort, and environmental impact. Vehicular networks, however, normally operate in communication-hostile environments and are characterized by dynamic topologies and volatile links, making it challenging to guarantee Quality of Service (QoS) and reliability of vehicular applications. To this end, the present work explores how the centralized coordination offered by Software-Defined Networking can improve the Quality of Service in vehicular networks, particularly for Vehicle-to-Vehicle (V2V) unicast routing and content dissemination. With regard to V2V routing, this work motivates the case for centralized network coordination by studying the performance of traditional MANET routing protocols when applied to urban VANETs, showing that they cannot provide satisfactory performance for modern vehicular applications because of their limited global network awareness, slow convergence, and high signaling. Hence, this work proposes and validates a centralized Multi-Flow Congestion-Aware Routing (MFCAR) algorithm to allocate multiple data flows on V2V routes. The first novelty of MFCAR is the SDN-based node-busyness estimation technique. The second novelty is the enhancement of the path-cost formulation as a linear combination of path length and path congestion, allowing the user application to fine-tune its QoS requirements between throughput and delay. Concerning content dissemination, this work proposes a Fairness- and Throughput-Enhanced Scheduling for Content Dissemination in VANETs (ROADNET), a centralized strategy to improve the tradeoff between data throughput and user fairness in deterministic vehicular content dissemination. ROADNET’s main novelties are the design of a graph-based multi-channel transmission scheduler and the enforcement of a transmission-priority policy that prevents user starvation. As additional contributions, the present work proposes a heuristic for the centralized selection of opportunistic content dissemination parameters and discusses the main security issues in Software-Defined Vehicular Networking (SDVN) along with possible countermeasures. The proposed techniques are evaluated in realistic scenarios (LuST), using discrete-event network simulators (OMNeT++) and microscopic vehicular-mobility simulators (SUMO). It is shown that MFCAR can improve PDR, throughput and delay for unicast V2V routing up to a five-fold gain over traditional algorithms. ROADNET can increase content dissemination throughput by 36% and user fairness by 6% compared to state-of-the-art techniques, balancing the load over multiple channels with a variance below 1%.
Research center :
Interdisciplinary Centre for Security, Reliability and Trust (SnT) > Networking Research Group (NetLab)
Disciplines :
Computer science
Author, co-author :
di Maio, Antonio ;  University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC)
Language :
English
Title :
Routing Strategies and Content Dissemination Techniques for Software-Defined Vehicular Networks
Defense date :
03 June 2020
Institution :
Unilu - University of Luxembourg, Esch-sur-Alzette, Luxembourg
Degree :
Docteur en Sciences de l'Ingénieur
Promotor :
President :
Secretary :
Jury member :
FnR Project :
FNR10487418 - Context And Content Aware Communications For Qos Support In Vanets, 2015 (01/05/2016-30/04/2019) - Thomas Engel
Funders :
FNR - Fonds National de la Recherche [LU]
Available on ORBilu :
since 24 June 2020

Statistics


Number of views
448 (21 by Unilu)
Number of downloads
0 (0 by Unilu)

Bibliography


Similar publications



Contact ORBilu