5G Next generation VANETs using SDN and Fog Computing Framework

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

5G Next generation VANETs using SDN and Fog Computing Framework ⇤

Ammara Anjum Khan⇤ , Mehran Abolhasan† , Wei Ni‡

Faculty of Engineering and Information Technology (FEIT), University of Technology, Sydney, Australia [email protected] † [email protected] ‡ [email protected]

Abstract—

The growth of technical revolution towards 5G Next generation networks is expected to meet various communication requirements of future Intelligent Transportation Systems (ITS). Motivated by the consumer needs for variety of ITS applications, bandwidth, high speed and ubiquity, researches are currently exploring different network architectures and techniques, which could be employed in Next generation ITS. To provide flexible network management, control and high resource utilization in Vehicular Ad-hoc Networks (VANETs) on large scale, a new hierarchical 5G Next generation VANET architecture is proposed. The key idea of this holistic architecture is to integrate the centralization and flexibility of Software Defined Networking (SDN) and Cloud-RAN (CRAN), with 5G communication technologies, to effectively allocate resources with a global view. Moreover, a fog computing framework (comprising of zones and clusters) has been proposed at the edge, to avoid frequent handovers between vehicles and RSUs. The transmission delay, throughput and control overhead on controller are analyzed and compared with other architectures. Simulation results indicate reduced transmission delay and minimized control overhead on controllers. Moreover, the throughput of proposed system is also improved.

Index Terms—Next generation VANETs, Software Defined VANETs, Fog computing, Edge Computing, 5G VANET architecture.

I. I NTRODUCTION VANETs have been regarded as key enabling technology of Next generation ITS, that is envisioned to offer a wide variety of versatile services to ITS consumers, ranging from transportation and road safety to infotainment applications like web browsing, video streaming file downloading and etc. [1]. The society of 2020 is predicted to be a connected society. The emerging idea of Internet of Things (IoT) together with intelligent and integrated sensor systems and in-home sensor networks are expected to potentiality exert influence on consumer’s daily lives and are expected to motivate huge market in near future. During the last few decades, VANETs are rapidly evolving. The number of connected vehicles is predicted to reach 250 million, by 2020 [2]. Moreover, by 2020, smarter and secure ITS are expected to be operational as a VANET cloud [3]. To enable smart vehicles to get connected,

978-1-5386-4790-5/18/$31.00 ©2018 IEEE

it is essential for a vehicle to access Internet and communicate with neighbours through wireless communication infrastructures. Nevertheless, current VANET architectures can not meet the latency requirements of future ITS applications in highly congested and mobile scenarios. The future trend of autonomous vehicles drives current VANET architectures, broaden their limits with hard real-time requirements. In addition, the maturity of cloud computing has adapted the invasion of vehicular space with cloud-based services. To provide an efficient communication and cooperation on large scale VANETs, millions of vehicles are widely spread in the environment of Internet of Vehicles (IoV), where drivers and passengers can enjoy all ITS services through Internet [4]. The cloudification of network resources through SDN and CRAN is another promising enabler for 5G Next generation vehicular networks. SDN is leading towards a revolutionary paradigm which controls the network in a centralized and programmable manner by decoupling the forwarding functions (data plane) and network controls (control plane). Moreover, due to its potential to offer flexibility, programmability and centralized knowledge, it facilitates flexible network management and control on large scale, with unified abstraction [4], [5], [6]. SDN has recently attracted research in mobile wireless networks like VANETs. SDN is also considered as one of the most promising techniques that can conveniently be applied to support the dynamic nature and dense deployment scenarios of future VANET functions and applications, while reducing the operational cost [7], [8]. On the other hand, the proposal of Cloud (or centralized) Radio Access Network (C-RAN) by China Mobile [9], is an affective and open cloud-based infrastructure that includes Baseband Units (BBUs), Remote Radio Units (RRUs), and antennas. The centralization of BBUs and cooperation among RRUs and antennas through open cloudbased infrastructure can effectively improve system spectrum efficiency. The virtualization techniques can share different resources operated by BBUs, according to service demands of different cells, thus reducing operational cost and power consumption [8]. On the other hand, the role of fog computing in IoT [10], [11] plays an important part in satisfying the real time service demands of future ITS scenarios. The main contributions of paper are as follows: • We propose a new hierarchical 5G Next generation

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

VANET architecture, motivated by the essence of SDN, C-RAN and fog computing technologies. • To support vehicles and end users with promptly responses, a new Fog Computing (FC) framework is proposed at the edge of network. The details of FC framework are discussed in section III. • The control functionality deployment of controller is divided in a hierarchical manner to reduce control overhead on centralised controller. Further details are discussed in section III . • The transmission delay, throughput and control overhead on controller are also analyzed and compared with other architectures. Simulation results reveal minimized transmission delay and control overhead on controller, considering different vehicle densities. • Moreover, the throughput of proposed architecture is also analyzed, using average bandwidth allocation scheme and adaptive bandwidth allocation scheme (i.e., by keeping in view different bandwidth demands of users). Simulation results reveal improved throughput as compared to throughput of traditional systems, and other proposed architecture. This paper is organized as follows. In section II, background and some related work is presented, to describe the motivations towards 5G enabler technologies for VANETs. Section III describes the topology and logical structure of architecture. In section IV, the performance of proposed architecture is analyzed and compared with other architectures. Finally, we conclude our work, and propose future directions. II. BACKGROUND AND R ELATED W ORK Due to high mobility and rapidly changing topology of VANETs, it is challenging to meet Next generation ITS services by using a single wireless infrastructure. Extensive research and efforts have been made from both industry and academia, in the field of Next generation Vehicular Communication Networks (VCNs), to get smart vehicles connected with the surrounding vehicles, road-side infrastructures, and internet through different wireless communication infrastructures [12], [13]. Therefore, Next generation vehicular networking is expected to adopt current cellular solutions like 3G, 4G Long Term Evolution (LTE), 3GPP, and is expected to be heterogeneous in terms of resources and network topology. LTE systems offer the benefits of large coverage, high throughput and low latency [14]. However, due to high vehicle mobility and dynamic network topology, it is relatively challenging to provide satisfied ITS services only through LTE systems. Integrating different access networks technologies like DSRC and LTE,is proposed to be a potential solution to meet various ITS service requirements [12]. However, building heterogeneous vehicular networks (integrating IEEE 802.11p with cellular technologies like 3G, 4G/LTE systems and etc.,), requires a deep understanding of heterogeneity and its associated challenges in VANETs. Due to high mobility and rapidly changing topology of VANETs, the handoffs among different wireless access infrastructures are more frequent, as compared

to traditional wireless networks thus causing service interruptions. There is a need to develop some unified ways to deal with control and management issues rising in heterogeneous VANETs on large scale. Furthermore, to provide consistency in services with the frequent topology changes and varying QoS demands in VANETs, the heterogeneous substrate cannot have a global view of all service requests, to make compromise and provide cooperation between all services. Inspite of all the efforts made in the field of heterogeneous VANETs, there is still a dramatic gap between the practical requirements of ITS services and what can be offered by existing heterogeneous VANETs. All of these issues above, call for rethink of the current network architecture for VANETs. Consequently, the research and development for the fifth generation (5G) systems have already been started [15], [16], [5], [6], [17], [18], [19], [20], [21]. On the other hand, SDN has been proposed as a promising technique that will play a key role in the design of 5G wireless communication networks [6]. SDN is proposed to be an effective technology to be capable of supporting the dynamic nature of VANETs and ITS applications, by facilitating flexible network management and optimization on large scale with unified abstraction [4]. In order to meet the demanding requirements of future ITS, SDN, Cloud Computing, Fog Computing are expected to be future candidate technologies for 5G VANETs. Some initial studies have also been carried out to integrate either of these technologies into Vehicular Communication Networks [4], [7], [22]- [23]. Nevertheless, the performance of SDN technology becomes limited in RSUs, when the number of vehicles connected with RSU increases [22]. The frequent handover problem in dense scenarios of VANETs, reduces the performance of SDN at RSUs [24]. However, it is also realised that the scalability of Wireless Distributed Networks (WDNs) is improved by using techniques like; clustering, multichannel routing and zoning [25] and [26]. Nowadays, C-RAN has been widely accepted to be a promising solution for heterogeneous networks [7]. In C-RAN, all RAN functionalities are performed in the centralized BBU pool, in cloud based infrastructure, which are connected to RRHs via fibre. The separation between the data plane and control plane via SDN can be built upon the open platform of Cloud-RAN by keeping in view the service demands of different users, thus reducing operational cost. In this paper, we propose a hierarchical 5G Next- generation VANET architecture by employing the concepts of SDN, C-RAN and fog computing as shown in Fig. 1. III. 5G N EXT GENERATION VANET A RCHITECTURE A. Topology Structure of Fog Computing (FC) Framework, C-RAN and SDN controller: To support vehicles and end users with promptly responses, FC framework is configured at the edge of network. FC framework is comprised of the following components. • Fog Computing-Zone Controllers (FC-ZCs): The FCZCs are the computing enhanced (i.e., CPU and storage)

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

SDN Controllers

FC-BBUC

Cloud-RAN

Distributed Control Plane

FC-BBUC

FC-BBUC - FC-ZC Communication

FC-BBUC Pool

FC-Vehicle – FC-Vehicle Communication FC-ZC – FC-CH Communication FC-Zone

FC-CH -FC-Vehicle Communication

FC-Zone 4G/LTE

FC-ZC

FC-CH -FC-CH Communication FC-ZC

WiMax

FC-Zone

FC-CH FC-CH FC-ZC FC-CH

3G

WiMax

FC-Vehicle

FC-CH FC-Vehicle

FC-Vehicle

FC-Vehicle

FC-CH

FC-CH FC-Vehicle

FC-Vehicle

Fig. 1: Topology Structure of 5G Next generation VANETs using SDN and Fog Computing (FC) Framework wireless access infrastructures such as, RSUs, Base Stations (BSs) connected with the BBU controllers, through broadband connections. In our case, a zone is defined as a group of vehicles that is registered with one RSU or a BS. Therefore, one FC-ZC is responsible for controlling one zone. Most of the data at edge is processed and saved by FC-ZCs. Moreover, FC-ZC devices are SDN- enabled, meaning they are under the control of SDN controller. The control overhead of devices (vehicles) remains in their own vicinity i.e., FC-zones or FC-Clusters, and is not sent to the SDN controller, unless required. Hence, FC-ZCs and FCzones play an important role in minimizing the overhead in the control plane. These devices act as both control pane and data plane elements. • Fog Computing-Cluster-heads (FC-CHs:) Further, FC-zones are divided into Fog Computing-Cluster heads (FC-CHs. Each FC-CH is controlled and managed by FC-ZC. FC-CHs are the vehicles, equipped with SDNenabled On Board Units (OBUs). • Fog Computing-Vehicles (FC-Vehicles): FC-Vehicles act as end users, and are also equipped with SDNenabled OBUs. The potential functionalities of OBUs include, packet forwarding, sensor localization system like Global Positioning system (GPS), power control, channel selection, interface selection and transmission mode (i.e, V2V or V2I communication). • Fog Computing BBU Controllers (FC-BBUCs): FCBBUC connects mutiple FC-ZCs with the backhual links. The FC-BBUC acts a as digital unit that is responsible for implementing the base station functionalities, from baseband processing to packet processing.

Several FC-BBUCs are placed in a central physical pool, to distribute FC-ZCs according to RF strategies. The advantage of using SDN-based virtualization for CRAN, in our proposed framework, is that the resource allocation and scheduling can be effectively and simply managed by the central controller, with a global view. Therefore, FC-BBUCs act as a bridge, connecting VANET infrastructure with the SDN controller. The FCBBUC collects the state information of different FCZCs connected with it, and by using its own local intelligence, it can make forwarding decisions, thus reducing the overhead on centralised controller. FC-ZC will communicate with the FC-BBUC, for inter FCzone communication. Therefore, FC-BBUCs are the data plane as well as control plane devices. • SDN Controller: As a core component, SDN controllers are responsible for network management and operations such as, rule generation, resource allocation and mobility management. Moreover, they can also perform some advanced network functionalities like, learning, network analysis and data pre-processing. In our case, SDN functionalities are distributed and also shared among local controllers i.e., FC-ZCs and FC-BBUCs and FC-CHs in a hierarchical manner. Moreover, SDN controllers is also responsible for Fog Orchestration and resource management of fog. • Optical Transmission Network (OTN): Optical Transmission Network (OTN) is responsible for transmitting and receiving digital signals to and from the FC-BBUC pool via optical fiber. The FC-ZCs send and receive digital signals to and from the FC-BBUC pool via optical fiber.

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

Application Layer

SDN Controller and C-RAN

Dynamic Offloading for HetVANETs

Access Control

Vehicular Ad hoc Network (VANET) Infrastructure Fog Computing Module/Inter-Zone Communication module

Fog Computing (FC)-Framework at Edge (Comprising of FC-BBUCs, FC-Zones and FCCHs )

Entertainment services

Controller Interface

Control Plane

Security Management and Authentication

Interface Compatibility Module

Mobility Management

Network Traffic Monitoring

Inter- FC-Zone Communication Module

Network Status monitoring Module

Network Status monitoring Module

SDN Cloud Computing Module

FC-BBUC FC-ZC FC-CH

SDN controller

Topology Information gathering Module

Data Plane

Fig. 2: Hierarchy of SDN controller, Cloud-RAN and Fog computing framework Communication Module

FC-BBUC Data gathering Module

FC-ZC

FC-CH FC-Vehicles

B. Logical Structure of proposed 5G Next generation VANET architecture: The logical structure of proposed architecture is divided into data plane, control plane and application plane as shown in Fig. 3. The data plane includes FC-Vehicles, FC-CHs, FCZCs and FC-BBUCs. Functionalities include data collection, quantization and then forwarding data to the control plane [27], [28]. The data plane devices can be configured in to the following function modules. • Information gathering module of FC-Vehicles, FCCHs, FC-ZCs, FC-BBUC: This module uses different sensors to record information related to position, speed and direction of vehicles and CCTVs, network cameras, lane checking cameras, etc. • Communication module of FC-Vehicles, FC-CHs, FCZCs, FC-BBUCs: This modules further includes V2V and V2I communication module. V2V provides wireless communication between two adjacent vehicles, that may be two FC-Vehicles or two FC-CHs or a FC-CH and a FC-Vehicle, by using WiFi/WAVE. V2I communication provides wireless communication between FC-CHs and FC-ZCs. Further, the communication module of FCBBUCs includes two types of communication modules, one is between FC-BBUC to FC-ZC and, other module is between FC-BBUC to SDN controller. Furthermore, inter-(FC-ZCs) and inter-(FC-BBUCs) communication is also performed by this module. The control level of SDN decides the flow rules or policy rules [29]. Since, we are using fog architecture at edge, therefore, SDN controller will operate in Hybrid Control Mode as shown in Fig. 1 and 2. The control plane includes SDN controllers, FC-BBUCs, FC-ZCs and FC-CHs. The FC-BBUC is the main control center or fog controller of fog framework. SDN controller functionalities are shared at the edge of network, between FC-BBUC, FC-ZCs and FC-CHs. SDN controller will not take full control of the network. Instead of sending specific flow rules, SDN controller will send abstract policy rule. The specific behaviour of policy rules will be decided by FCZCs, FC-BBUCs and FC-CHs, depending on their own local intelligence [30]. Following are the control plane function modules. • Information gathering modules of FC-BBUC, FC-ZCs

Fig. 3: Logical Structure of proposed 5G Next generation VANETs and FC-CHs: To draw global view map of network, based on data information provided by the data plane. • Computing and Storage modules: These modules are deployed in fog computing framework devices and cloud computing centres. • Network status monitoring module: Responsible for monitoring the links of 5G SDN-based VANET architecture. • Inter-FC-zone communication module: Configured in FC-ZCs to provide inter zone communication in fog network. • Inter-FC-BBUC communication module: Configured in FC-BBUCs to provide inter BBU communication in Cloud-RAN. • Intra-FC-zone communication module: Configured in FC-CHs to provide communication between FC-CHs within a FC-Zone. The application plane is responsible for generating rules and strategies, based on different application requirements of users/vehicles, and forward these rules to the control plane. Details are in Fig. 3. Table I shows some of the requirements of proposed architecture using Cloud computing (SDN controller cloud and C-RAN) and fog computing framework. IV. C OMPARISON OF T HROUGHPUT, T RANSMISSION DELAY AND C ONTROL OVERHEAD ON CONTROLLERS The performance of our proposed architecture is analyzed and compared with two architectures i.e., traditional architecture and 5G VANET architecture proposed in [31], named as 5G SD VANETs for our comparison. In traditional architecture, every vehicle communicates with the RSU directly, whereas in [31], each node has to send signalling information to a node closer to RSU. The performance of proposed architecture is investigated, by analysing the throughput, transmission delay, and control overhead on controllers, using MATLAB. Considering the real bandwidth requirements of vehicles, an adaptive bandwidth allocation scheme is also used to optimize

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

TABLE I: Requirements of Proposed architecture Requirements Mobility Support Geographical Distribution Security Location of Server Nodes Distance between vehicle and servers Location Awareness Delay Control functionality deployment Controller Operation Mode

Cloud Computing framework (C-RAN and SDN controller) Limited Centralised and Distributed Undefined Within internet

Fog Computing Framework

multiple hops

single hop/very few hops Yes Very Low Hierarchical

Supported Centralised and Distributed Can be Defined Edge Network

No High Hierarchical Hybrid (shared between Fog computing devices and SDN controller)

Hybrid (shared between zone controllers and CHs)

the throughput of fog framework. Simulation results in Fig. 4 show improved throughput, as compared to throughput in [31], and throughput of traditional architecture. It is shown in Fig. 5 that the throughput of a FC framework using both average and adaptive bandwidth allocation scheme is improved as compared to throughput of fog cell in [31]. We analyse and 800

600

0.14 Traditional Architecture Proposed 5G SD VANETs [31]

500 0.12

400

Traditional Architecture Proposed 5G SD VANETs[31]

300

0.1 Delay (ms)

Throughput (Mbps)

700

compare the transmission delay of vehicle, in fog framework, considering different vehicle densities. In [31], as the complexity of handovers between vehicles and RSU is increased, with an increase in multihop relay vehicles, the propagation delay increases. Using the concept of zones and clusters, the number of multihop relay vehicles is reduced, thus reducing delay. For analysis, we use ALM [32], as a clustering strategy. Fig. 6 shows, there exist a minimum transmission delay of 0.06ms, as compared to transmission delay of traditional architecture and 5G Software Defined Vehicular Networks architecture [31]. The reason is, in our proposed FC-Framework, the control functionalities are divided among different controllers, as discussed earlier in section III, and data, processing and applications are concentrated in devices/vehicles at the network edge, rather than existing almost entirely in the cloud. Moreover, devices/vehicles communicate peer-to-peer to efficiently share/store data and take local decisions, thus reducing delay. Another reason is that due to more than one FC-CHs within FC-zones, the FC-CHs are directly communicating with the RSUs, thus reducing number of relay hops for transmission and reducing delay. It is also seen that, when the density of vehicles is low, the distance among adjacent vehicles is far away, therefore, the success transmission probability of link is low, thus, delay will be increased. Increasing vehicle density, will decrease the distance among adjacent vehicles and therefore, the success of transmission probability will be increased. Therefore, delay will be minimized. We also

200 100 0 10

15

20

25

30

35

40

45

50

55

60

0.08 0.06 0.04

No. of Vehicles 0.02

Fig. 4: Throughput Comparison

0 10

15

20

25

30

35

40

45

50

55

60

Throughput (Mbps)

No. of Vehicles 1000

Fig. 6: Delay Comparison

800

analyse the control overhead on controllers. Fig. 7 shows that the control overhead on controller is significantly reduced as compared to the control overhead on controller using traditional architecture and 5G Software Defined Vehicular Networks architecture in [31]. This is due to hierarchical distribution of controllers in control plane, and practical use of zones and clusters in our proposed FC-framework.

600

400 Average Bw (Proposed) Adaptive Bw (Proposed) Average Bw (5G SD VANETs [31]) Average Bw (5G SD VANETs [31])

200

0 10

V. C ONCLUSION AND FUTURE D IRECTIONS 15

20

25

30

35

40

45

50

55

60

No. of Vehicles

Fig. 5: Comparison of Throughput using average and adaptive bandwidth allocation schemes

We propose a new hierarchical 5G Next generation VANET architecture, by employing the concepts of SDN, C-RAN and fog computing technologies. Moreover, a new Fog Computing framework is proposed at edge of network.The distributed

2018 15th IEEE Annual Consumer Communications & Networking Conference (CCNC)

Control Overhead (No. of packets)

250

200

150

100 Traditional Architecture Proposed 5G SD VANETs [31]

50

0 10

15

20

25

30

35

40

45

50

55

60

No. of Vehicles

Fig. 7: Comparison of Control overhead on controller support of fog computing framework, offers delay-sensitive, location-awareness and mobility-based real time services suitable for future ITS scenarios. Using SDN and C-RAN technologies, the proposed architecture provides flexibility, programmability and effective resource allocation using control plane and centralised global knowledge, thus leading towards significant reductions in operating cost of operators. Simulation results reveal improved throughput, reduced transmission delay and minimized overhead on controllers. In future, we aim to design analytical models for optimum resource sharing in C-RAN using a globalized view. Future challenges include, optimizing route selection, designing protocols at SDN controller for load balancing, improving service efficiency provision, due to massive traffic increase for 5G Next generation VANETs. R EFERENCES [1] P. Papadimitratos, A. De La Fortelle, K. Evenssen, R. Brignolo, and S. Cosenza, “Vehicular communication systems: Enabling technologies, applications, and future outlook on intelligent transportation,” IEEE Communications Magazine, vol. 47, no. 11, pp. 84–95, 2009. [2] “Gartner, "connected cars from a major element of internet of things" http://www.gartner.com/newsroom/id/2970017,” 2015. [3] S. Jia, S. Hao, X. Gu, and L. Zhang, “Analyzing and relieving the impact of fcd traffic in lte-vanet heterogeneous network,” in Telecommunications (ICT), 2014 21st International Conference on. IEEE, 2014, pp. 88–92. [4] C. Jiacheng, Z. Haibo, Z. Ning, Y. Peng, G. Lin, and S. Xuemin, “Software defined internet of vehicles: architecture, challenges and solutions,” Journal of Communications and Information Networks, vol. 1, no. 1, pp. 14–26, 2016. [5] S. Sun, M. Kadoch, L. Gong, and B. Rong, “Integrating network function virtualization with sdr and sdn for 4g/5g networks,” IEEE Network, vol. 29, no. 3, pp. 54–59, 2015. [6] M. Zolanvari, “Sdn for 5g.” [7] K. Zheng, L. Hou, H. Meng, Q. Zheng, N. Lu, and L. Lei, “Soft-defined heterogeneous vehicular network: architecture and challenges,” arXiv preprint arXiv:1510.06579, 2015. [8] B. Cao, Y. Li, C. Wang, G. Feng, S. Qin, and Y. Zhou, “Resource allocation in software defined wireless networks,” IEEE NETWORK, vol. 31, no. 1, pp. 44–51, 2017. [9] C. Mobile, “C-ran: the road towards green ran,” White Paper, ver, vol. 2, 2011. [10] F. Bonomi, R. Milito, J. Zhu, and S. Addepalli, “Fog computing and its role in the internet of things,” in Proceedings of the first edition of the MCC workshop on Mobile cloud computing. ACM, 2012, pp. 13–16. [11] I. Stojmenovic and S. Wen, “The fog computing paradigm: Scenarios and security issues,” in Computer Science and Information Systems (FedCSIS), 2014 Federated Conference on. IEEE, 2014, pp. 1–8.

[12] K. Zheng, Q. Zheng, P. Chatzimisios, W. Xiang, and Y. Zhou, “Heterogeneous vehicular networking: A survey on architecture, challenges, and solutions,” IEEE Communications Surveys & Tutorials, vol. 17, no. 4, pp. 2377–2396, 2015. [13] Z. He, D. Zhang, and J. Liang, “Cost-efficient heterogeneous data transmission in software defined vehicular networks,” in High Performance Computing and Communications (HPCC), 2015 IEEE 7th International Symposium on Cyberspace Safety and Security (CSS), 2015 IEEE 12th International Conferen on Embedded Software and Systems (ICESS), 2015 IEEE 17th International Conference on. IEEE, 2015, pp. 666– 671. [14] G. Araniti, C. Campolo, M. Condoluci, A. Iera, and A. Molinaro, “Lte for vehicular networking: a survey,” IEEE Communications Magazine, vol. 51, no. 5, pp. 148–157, 2013. [15] M. Agiwal, A. Roy, and N. Saxena, “Next generation 5g wireless networks: A comprehensive survey.” [16] A. Gohil, H. Modi, and S. K. Patel, “5g technology of mobile communication: A survey,” in Intelligent Systems and Signal Processing (ISSP), 2013 International Conference on. IEEE, 2013, pp. 288–292. [17] Z. Ma, Z. Zhang, Z. Ding, P. Fan, and H. Li, “Key techniques for 5g wireless communications: network architecture, physical layer, and mac layer perspectives,” Science China Information Sciences, vol. 58, no. 4, pp. 1–20, 2015. [18] S. Chen, F. Qin, B. Hu, X. Li, and Z. Chen, “User-centric ultradense networks for 5g: challenges, methodologies, and directions,” IEEE Wireless Communications, vol. 23, no. 2, pp. 78–85, 2016. [19] M. X. Gong, R. Stacey, D. Akhmetov, and S. Mao, “A directional csma/ca protocol for mmwave wireless pans,” in Wireless Communications and Networking Conference (WCNC), 2010 IEEE. IEEE, 2010, pp. 1–6. [20] A. El-Keyi, T. ElBatt, F. Bai, and C. Saraydar, “Mimo vanets: Research challenges and opportunities,” in Computing, Networking and Communications (ICNC), 2012 International Conference on. IEEE, 2012, pp. 670–676. [21] R. N. Mitra and D. P. Agrawal, “5g mobile technology: A survey,” ICT Express, vol. 1, no. 3, pp. 132–137, 2015. [22] K. Liu, J. K. Ng, V. C. Lee, S. H. Son, and I. Stojmenovic, “Cooperative data scheduling in hybrid vehicular ad hoc networks: Vanet as a software defined network,” IEEE/ACM transactions on networking, vol. 24, no. 3, pp. 1759–1773, 2016. [23] N. Zhang, S. Zhang, P. Yang, O. Alhussein, W. Zhuang et al., “Software defined space-air-ground integrated vehicular networks: Challenges and solutions,” arXiv preprint arXiv:1703.02664, 2017. [24] T. Taleb and K. B. Letaief, “A cooperative diversity based handoff management scheme,” IEEE Transactions on Wireless Communications, vol. 9, no. 4, 2010. [25] X. Li, P. Djukic, and H. Zhang, “Zoning for hierarchical network optimization in software defined networks,” in Network Operations and Management Symposium (NOMS), 2014 IEEE. IEEE, 2014, pp. 1–8. [26] M. Abolhasan, T. Wysocki, and E. Dutkiewicz, “A review of routing protocols for mobile ad hoc networks,” Ad hoc networks, vol. 2, no. 1, pp. 1–22, 2004. [27] M. Feng, S. Mao, and T. Jiang, “Enhancing the performance of futurewireless networks with software-defined networking,” Frontiers of Information Technology & Electronic Engineering, vol. 17, no. 7, pp. 606–619, 2016. [28] I. Ku, Y. Lu, M. Gerla, F. Ongaro, R. L. Gomes, and E. Cerqueira, “Towards software-defined vanet: Architecture and services,” in Ad Hoc Networking Workshop (MED-HOC-NET), 2014 13th Annual Mediterranean. IEEE, 2014, pp. 103–110. [29] M. Mendonca, K. Obraczka, and T. Turletti, “The case for softwaredefined networking in heterogeneous networked environments,” in Proceedings of the 2012 ACM conference on CoNEXT student workshop. ACM, 2012, pp. 59–60. [30] N. B. Truong, G. M. Lee, and Y. Ghamri-Doudane, “Software defined networking-based vehicular adhoc network with fog computing,” in Integrated Network Management (IM), 2015 IFIP/IEEE International Symposium on. IEEE, 2015, pp. 1202–1207. [31] X. Ge, Z. Li, and S. Li, “5g software defined vehicular networks,” arXiv preprint arXiv:1702.03675, 2017. [32] E. Souza, I. Nikolaidis, and P. Gburzynski, “A new aggregate local mobility (alm) clustering algorithm for vanets,” in Communications (ICC), 2010 IEEE International Conference on. IEEE, 2010, pp. 1–5.