SDN-enabled vehicular networks: Theory and practice within platooning applications

Abstract

With the recent developments of communication technologies surrounding vehicles, we witness the simultaneous availability of multiple onboard communication interfaces on vehicles. While most of the current interfaces already include Bluetooth, WiFi, and LTE, they are augmented further by IEEE 802.11p and the 5G interfaces, which will serve for safety, maintenance, and infotainment applications. However, dynamic management of interfaces depending on application becomes a significant issue that can be best addressed by Software Defined Networking (SDN) capabilities. While SDN-based vehicular networks have been promoted previously, none of these works deal with practical challenges. In this paper, we propose and develop a practical framework that realizes SDN-based vehicular networks for a wide range of applications. Through this framework, we demonstrate a truck platooning application as a use case in which the two truck platoons strive to merge and establish connectivity. The route from source vehicle (i.e., Platoon Leader A) to destination (i.e., Platoon Leader B) is computed with the help of the SDN Controller to transmit the Beacon Safety Messages through Road Side Units (RSUs) at the MAC layer without relying on IP for proper platooning operations. The results show the efficiency of the SDN-based approach compared to the traditional routing approaches.

Introduction

The number of vehicles on the roads has risen significantly with the increasing mobility of the people, and goods [1], which led to many research studies on traffic congestion, safety, and other transportation problems. However, with the emerging developments on vehicles, IoT, and communication technologies, vehicles are now moving to a new era where they are touted as the next personal smart devices which will have a tremendous impact on the quality of our lives through the services offered by their availability such as transportation, traffic congestion, safety, infotainment, and environmental crowd sensing. Added to this reality is the concept of self-driving or autonomous drivers, which will come with their own capabilities to disrupt the transportation industry in the next decades [2]. Thus, we will essentially see smart vehicles emerging as IoT-like devices that are always connected while being aware of what is happening in the surroundings through its sensors and acting accordingly.

The challenges surrounding the vehicles for the realization of the aforementioned services are not new. Over the last two decades, there have been many efforts to alleviate the problem of traffic safety due to increasing number of vehicles such as road information signs, radio communication for hazards for the drivers in addition to vehicle-to-vehicle communication (V2V) technologies. For enabling vehicles to communicate with each other (i.e., V2V) and with infrastructure (i.e., vehicle-to-infrastructure, V2I) or in more generic term vehicles to everything (i.e., V2X), there have been tremendous standardization efforts [3]. For instance, there are safety applications that are based on short-distance broadcasting messages to the neighbor vehicles to prevent accidents on the road. In order to satisfy these demands, the IEEE 802.11p standard, also known as DSRC, has been developed as standard network technology for V2V communication [4]. The IEEE 802.11p is the oldest technology to broadcast basic safety messages between vehicles and/or roadside units (RSUs). Despite the mandate for DSRC from the US government, the V2X standard as part of 3GPP specifications is picking up much faster [5], especially with the rolling of 5G technology. Therefore, using both IEEE 802.11p and 4G/LTE/5G would complement each other, and thus they will be able to co-exist to address different needs. For instance, V2X may be more suitable for long distance communications while DSRC can still exist for enabling V2V sidelinks. There are also many developments regarding the communications within a vehicle as a Cyber-Physical System (CPS) [6]. For instance, the sensors within a vehicle, as well as other control units, can talk to each other using Ethernet-like technologies (i.e., CAN Bus [7]) or Bluetooth. This communication requires extra radio interfaces that should be available on modern vehicles through their on-board units (OBUs) which may host all of these diverse interfaces.

Nevertheless, exploiting such a diverse set of radio access for different applications, stakeholders, and environments for vehicles brings many challenges. For instance, the question of deciding which technology to use for each traffic type arises as one of the primary challenges [8]. In addition, this decision may need to be managed by external entities for the sake of application needs rather than leaving it merely to the individual vehicle's needs. This is where Software-Defined Networking (SDN) [9] comes into play which is a great fit to remotely control traffic passing through the vehicles' OBUs and RSUs. SDN divides the control and data plane of the communication while providing central access to the network switches through an SDN Controller. SDN-based vehicular networks (VANETs) can provide a smooth transition from one radio access to another by considering different metrics such as the type of application, location of the vehicles, and the density of the traffic in a given radio interface. SDN can also facilitate multi-hop broadcast messages if one of the safety messages is considered essential and useful for dissemination to further distances.

While SDN-based VANETs have been proposed for different use cases in the past [10], there is an important item missing in such studies: None of the ideas were implemented and tested in a realistic environment to understand the engineering challenges at scale under different communications standards. Therefore, in this paper, we first propose a generic framework that offers the use of multiple radio access technologies for each vehicle and RSU with an SDN-based switch and controller. Our framework proposes using 802.11p for safety applications, 5G for infotainment and self-driving purposes, and Bluetooth for in-vehicle communications. Due to its wide coverage and availability, the SDN control channel utilizes LTE as a separate interface. This is a comprehensive framework that can be customized based on the resources and needs of the vehicles.

As a second contribution, we demonstrate the applicability of this framework within a practical use case, namely platooning to address one of its challenges. Recently, there has been a great effort in enabling platooning for a variety of vehicles, including trucks.¹ A critical issue in platooning is that even though the platoons can be arranged in advance, these platoons might get separated due to the roads' dynamic characteristics, or there can be new trucks that need to join another platoon. This problem has not been studied and SDN is a perfect use case to tackle this problem. Specifically, given that platoons would be managed by the same company, they can employ their own SDN controller. Thus, we propose utilizing our SDN-based framework to tackle the connectivity establishment problem in platooning through the existing BSM messages. Specifically, we exploit multi-hop BSM messaging, which would enable two different platoons to talk and merge efficiently. SDN allows us to address this connectivity establishment within the MAC layer without resorting to IP, which is not only faster but also does not bring any additional changes to the existing DSRC stack.

We developed our framework by implementing a module within the NS-3 simulation environment² to test its feasibility and effectiveness, which not only supports the underlying technologies but also allows access from the SDN controller. Specifically, we extended the Openflow module [11] for NS-3 simulator to integrate 802.11p, LTE and 5G [12]. In addition to the default Ethernet ports in our scenarios. To the best of our knowledge, this is the first realistic and comprehensive implementation of SDN-based vehicular networks that can serve various needs. Under this framework implementation, we compared the proposed SDN-based platoon merge approach to other existing solutions such as Ad-hoc On-demand Distance Vector (AODV) [13] routing protocol. The experimental results show that SDN has a significant advantage in accommodating platoon merge operations quickly and efficiently.

The rest of this paper is organized as follows: In Section 2, we explain some related work and give some background information in 3. Section 4 introduces our proposed framework and Section 5 explains our specific solution to Platooning. Section 6 presents our experimental evaluations of our framework. We conclude the paper in Section 7.