20.3.1 Fog Computing

Fog computing is an emerging technology that extends the cloud capabilities, such as computation, storage, and networking services from the remote data center to closer to the end-user [16–18]. By locally processing the requests, data traffic is reduced on the core network, thereby reducing the processing time, latency, and network cost. The fog computing architecture is implemented on multiple layers of a network's technology [19, 20]. By adding layers of fog nodes, applications can be partitioned to run at the optimal network level [21, 22]. Fog computing specifically aims the time-critical applications, e.g. real-time gaming, virtual reality, emergency response system, vehicular networks, real-time health care, Internet of Things (IoT), and so on, to name a few [23, 24].

20.3.2 The Need of Fog Computing in 5G Networks

The future 5G networks aim for ultra-low latency and high bandwidth. 5G communication will be driven by high-frequency millimeter waves, and a 5G cellular network will connect mobile users to the base station (BS) connected to the core network. The low latency radio interfaces in the 5G network aim to reduce the delays between a user and a base station. However, the forwarding of user application requests from core network to the cloud would significantly increase latency. Consequently, 5G networks require moving processing power closer to the end-user, which can be accomplished with fog computing. Fog computing can be enabled in 5G by leveraging both the small and macro-based stations to achieve the ultra-low latency demanded in 5G networks. A fog computing–based radio access network (F-RAN) could provide 5G with reduced latency and high energy efficiency.

20.4.1 Cellular Architecture

In 5G, each device (cell phone) has a constant “home IP address,” in additional to a “care-of” address. Care-of address represents actual current location whereas IP identifies network interface and location addressing. Before communicating with a mobile, a computer passes a message packet to its IP address. Then, it is moved to the care-of address by directory server through a tunnel. The directory tunnel replies back to the computer and informs it about the correct address, and then further communication starts. As compared to IPV4 addressing, a connectionless protocol is being used in the switched layer. A total of 128 bits are used, which are subdivided into four categories, dedicated for different functions. The first 32 bits are used for the “home” IP address, and the other 32 are specifically used for the “care-of” address. The second part is for communication among cell phones and other devices (e.g. computer). The third part is used for tunneling to establish interconnection among wired and wireless networks. In the third part, the directory server uses the IP address of the mobile phone to create a communication channel. The last 32 bits of IPv6 address may be used for local addressing, e.g. for VPN (virtual private networks) sharing [25]. Figure 20.2 shows a general 5G cellular network architecture [10].

5G phone is designed to be an open source and includes every layer from physical layer to the end-user layer (application layer). Open wireless architecture, which was proposed for 4G mobiles phone, is adapted in this conceptual phone, and some changes in network layer and other layers up to application layer are made. The proposed concept divides the network layer into further two sublayers: upper and lower network layers. The lower layer is for each interface and the upper layer is for mobile terminal. The middleware between the upper and lower layer translates the IP addresses [26]. The protocol stack for the proposed 5G mobile phone concept is illustrated in Table 20.2. To allow the maximum optimized usage of specific wireless implementations of transport protocols, the open transport layer is proposed in this concept. The proposed concept will open a way to select many wireless technologies on different QoS and price constraints for different services, e.g. voice over Internet protocol (VoIP), streaming Internet protocol television (IPTV), web, messaging, and gaming. To fulfill this, 5G mobiles will create a database that will gather and provide information for all the services and technologies in the cellular device. Development modes will provide the best QoS and cost-to-use available wireless technologies within a 5G phone. Figure 20.3 explains the model, proposing a 5G architecture that is all-IP based and mobile network interoperability composed of nodes and independent radio access technologies (RATs). It is anticipated that 5G will change the concept of station-based-centric architecture of mobile networks, and methods of uplink/downlink control and data channels, to make routing and the flow of information in a better way. Currently, mobile systems are not letting devices communicate with each other directly. Every communication works through a base station. Tehrani et al. proposed a two-tier cellular network that allows a base-to-device communication, and device-to-device communication, as well [27]. In a two-tier mobile system, as users' data is passed by other users' cell phones, safety and security are the biggest concerns. To ensure breaches do not occur, the systems must be designed with smart interface strategies and managed allocation of resources. The traditional macro-cell tier includes a base station for device communication. In 5G-based device-to-device (D2D) communication, the device is directly connected to another device without a base station and transmission is directly among the two nodes with the assistance of other nodes. Base stations are still there, but devices will be allowed to create an ad-hoc mesh network. Another important thing is pricing and charge of D2D services. Pricing models are needed to be designed to let devices to participate in D2D communication [11].

20.4.2 Energy Efficiency

The main goal of 5G is to reduce energy. Therefore, energy efficient infrastructures play a vital role in 5G networks. Rowell et al. consider a usage of combined spectral efficiency and optimization of energy [28]. In a user-centric approach, a user's equipment is able to select uplink and downlink channels from various BS, depending upon amount of load, condition of channel, applications, and services requirements. Direct association cannot be performed entirely with user's equipment (UEs); it must consider both the channels at the time when they make an association with base stations (BSs). In this way, demodulation of signals and data is beneficial to save energy. An example of this scenario is a mobile base station (MBS), which may initiate signaling of a BS, and a server base station (SBS) to serve all data requests. If there is no data request for that time duration, we can hibernate the SBS. Zhang et al. proposed the same technique for decoupling of data in those UEs that are connected to SBS, according to the functionality provided by MBS, and output the result, which ensures less energy consumption and a lower chances of interference [29]. Hu et al. proposed C-RAN architecture for energy efficiency in a manner that UEs and Remote Radio Head (RHHs) serve almost the same number of times [30]. They also proposed an approach to interface management to decrease power consumption of SBSs and MBSs.

20.4.3 Two-Tier Architecture

Another proposed architecture for 5G is two-tier architecture. In two-tier architecture, MBS works at a higher tier and SBSs work under MBS, known as the lower tier. Higher and lower tiers share the same frequency band. There exists a macro cell, which covers all the small cells. These small cells include micro, pico, and femto cells. Small cells enhance the services and coverage area of macro cells. Two-tier architecture is enhanced to multitier architecture by using D2D and cognitive radio network (CRN)–based communication.

To separate the indoor and outdoor users is a big challenge. Wang et al. proposed a solution to this challenge by enhancing the MBS's responsibilities [14]. They deployed a huge array of antennas, having some of its elements distributed across macro cells, and they are connected by fiber optics to MBS. For communicating with MBS, a large array of antennas and SBS are deployed in buildings. SBS or Wi-Fi connects UEs with all other UEs in a building.

Wang et al. also proposed that a mobile (small-cell) placed inside a bus wants to communicate with internal UEs and an antenna that is located outside of the bus with MBS [14]. Thus, inside the bus, all UEs appear as a signal group or a single unit with corresponding MBS. SBS appear there as an MBS for all UEs inside the bus.

There are two basic issues while deploying two-tier architecture in small cells. Interference management and Backhaul data transfer. The interference may include inter-tier and intra-tier interference. The inter-tier interference will be as a result from MBS to SBS and from MBS to SBS's UEs. In intra-tier interference, resultant interference will be from SBS's to other SBS's UEs. In backhaul data transfer, an opaque placement of small cells in a network requires large data transfer rate and cost-efficient architecture. There are different models proposed by Nisha et al. for backhaul data transfer. They are wireless point-to-mMultipoint (PTMP), wired optical fiber (WPF), and wireless point-to-point (PTP). WPF backhaul data transfer is modeled through establishing the web objects framework (WOF) link from every BSS to MBS. It is an expensive and time-consuming process. PTMP deploys a wireless PTMP-base station on MBS. MBS then communicates with SBS and transfers backhauled data to the core network. In PTP, antennas are deployed in line-of-sight and have been proved to supply data at a high rate and capacity at low cost.

20.4.4 Cognitive Radio

A CRN is based on a two-tier architecture. It considers either MBS or SBS to implement cognitive radio properties at different mediums. Xuemin et al. [31] mention types of CRN-based architecture as (1) cooperative CRNs and (2) noncooperative CRNs. Cooperative CRNs use only cognitive and licensed bands/channels due to QoS. A licensed channel is used to serve UEs by a SBS, and the opportunistic access to the licensed channel is used to transfer backhauled data to the MBS. A noncooperative approach establishes a multi-RAT system, including radio interface and operates at temporary and licensed channels by programmable unit (PU), known as cognitive channel (CC). The source unit (SU) runs only on CC to form a CRN. It overlaps on an existing cellular wireless network. As there is connection between two networks, they can be integrated in the upper layer and must separate in the physical layer. The end-users use a cognitive and licensed channel to connect with a nearer MBS, and the communication between SBS and MBS occurs through a licensed channel.

In 5G, SBS works as an SU to monitor all activities and work on temporarily used frequency bands of small cells [31]. This band is called spectrum hole [32]. The PU provides services by micro cell to their UEs while having a minimum disturbance to macro cell functions. For the small cells, a dynamic pricing model based on theoretical framework is suggested in [28]. Due to this model, high data rates are achieved by UEs as compared to macro-cell and UEs pricing model is different as compared to macro-cell. Multitier architecture consists of different type of small cells, broadcasting, and serving D2D communication to serve users with QoS that are required for spectrum efficiency and energy efficient manners [33].

20.4.5 Cloud-Based Architecture

McKeown et al. defines cloud computing as an infrastructure providing simple, on demand, scalable, and less costly access to configurable resources that are at shared pool without worrying about resource management [34]. Mobile cellular communication and cloud's integration can provide tremendous beneficial growth to communication system. Here we briefly discuss cloud-based radio odes (C-RANs) for 5G as mentioned by [35]. Beyond C-RAN, the aim is to execute MBS functions in cloud hence; MBS's functionality is divided into two layers: control layer and data layer. Data layer contains heterogeneous physical resource and perform signal-processing task. Control layer performs resource management and baseband processing. C-RAN provides dynamic service without costly charges to network devices. Two basic components of MBS are baseband unit (BBU) and RRH. BBU implements baseband processing by utilizing baseband processor. Whereas, RRH is used for performing radio functionalities. C-RAN provides scalable and flexible architecture due to most of the BBUs placed in clouds, and RRHs are in MBSs.

Checko et al. proposed two-layered C-RAN architecture, in which two C-RAN are used based on MBSs [35]. The MBSs functionalities are divided into two groups: full centralized C-RAN and partially centralized C-RAN. In full-centralized C-RAN, BBU, and all high-level core functionalities of MBS are placed at the cloud level. While RHH and some other core functionalities are located in MBS, the high-level functionalities are still in the cloud, as mentioned in Figure 20.4.

20.5.1 HetNet

Humans use services such as social networking, VoIP, video conferencing, video streaming, and multiplayer gaming. Such forms of data communication are called human-type communication (HTC). When machines communicate with each other without human interference, then this form of data communication is called machine-type communication (MTC). When humans and machines communicate with each other, then this form of communication is termed as combined-type communication (CTC). CTC and MTC are important for IoT and cyber physical systems (CPS). 5G networks are presented in different use cases to establish the aforementioned communication mechanisms. The use cases are categorized in three ways: HTC, MTC, and CTC.

HTC Use Case

1. Inelastic. This use case is to satisfy user's demand for high-definition multimedia content at anytime, anywhere with a range of mobility and communication standards.
2. Elastic. This handles the large number of users and availability of Internet anytime, anywhere with the mobility communication standard.

MTC Use Case

1. Inelastic:
   1. Intelligent roads: A set of sensor and video cameras are arranged along a network of roads.
   2. Inter-vehicular communication. Autonomous driving, intelligent intersections, and traffic management need inter-vehicular communications.
   3. Industrial process automation. Sensors monitor automated industrial processes, and report their readings to a controller.
   4. Theft control. Sensor-based systems can be used to detect product theft in supply chain and/or in stores.
2. Elastic:
   1. Home automation. Different devices inside a home collaborate to accomplish a task.
   2. Intelligent parking. Sensors attached to parking spaces can communicate with online servers to provide information about available parking spaces in a particular area.

CTC Use Case

1. Inelastic. Critical control of remote devices and remote network management.
2. Elastic. Controlling and monitoring different home appliances from within a house or over the Internet.

5G research community still face issues, such as a scalable network architecture, QoS framework, intelligent network selection, and inter-network handover algorithms to satisfy diverse HTC, MTC, and CTC requirements. 5G supports a large amount of data initiated by HTC, MTC, and CTC [37].

20.5.2 Beam Division Multiple Access (BDMA)

5G technology uses CDMA and beam division multiple access (BDMA). BDMA allows data rate up to 100 Mbps with high mobility and at low mobility it is up to 1 Gbps. It uses an orthogonal beam, which is allocated to each mobile station. To increase the capacity of the system, an antenna beam is used that divides the band allocated to the base station. Mobile stations and the base stations are in the line of sight; both transmit the beam for the proper communication and reducing interference [38]. BDMA is shown in Figure 20.5.

20.5.3 Mixed Bandwidth Data Path

The architecture of 5G combines pico cell, micro cell, and macro cell, and can utilize bandwidth based on WLAN and CDMA2000. The WLAN network covers a small area, whereas CDMA2000 covers a large area. Moreover, in 5G, data requests can be sent from one network (MN, BS, PDSN, CN) and replies can arrive from the other network (CN, PDSN, AP, MN). The bandwidth selection is made with the exchange of messages between the networks, as depicted in Figure 20.6 [39].

20.5.4 Wireless Virtualization

5G supports network virtualization. Virtualizing mobile networks and sharing of their resources will provide more efficient utilization of wireless networks. Network virtualization can reduce the amount of base station equipment, as well as reduce energy consumption [40].

20.5.5 Flexible Duplex

5G allows very flexible and dynamic test-driven design (TDD) transmission resource. Existing TDD-based mobile technologies include TDD-LTE for restriction on uplink/downlink configuration. The TDD requires interference like direct base-station-to-base-station and device-to-device that is similar to base-station-to-device and device-to-base-station [41].

20.5.6 Multiple-Input Multiple-Output (MIMO)

Multiple-input multiple-output (MIMO): Multiple antennas are used to multiplex messages for multiple nodes, keeping the energy radiations toward the specific directions, which decreases the mobile interference. 5G systems do not encourage the idea of keeping complete control at the infrastructure side and take advantage of intelligence at the device side among the different layers of the stack protocol. The 5G idea requires an enhancement at the device level (component change); it also has implications at the architectural level. As wireless technology is emerging and data traffic is rapidly growing, the demand for swift and reliable base stations has increased. A new term, massive MIMO, describes a system that is emerging to fulfill this need. There are many benefits of this technique that are being realized by using large antenna arrays at the transmitters of receivers [2]. Massive MIMO and advanced code technology is a new concept in wireless paradigm for multi-users with multi-antenna systems [42]. It is able to provide better improvement in radiated energy efficiency and bandwidth. Advanced code refers to polar code, which are the first-ever codes that achieved the capacity of symmetric binary-input discrete memory-less channels (B-DMCs). They are regarded as the most promising code candidates for 5G, due to having advantages in performance, complexity, and flexibility. A MIMO has a number of receivers and transmitters and channels by which they communicate. Massive MIMO with deployment of multiple antennas is represented in Figure 20.7.

20.5.7 M2M

In 5G, M2M includes three basic requirements: Support for a large number of devices with low rate of transfer, sustenance of low data rate in non-real-time situations, and lesser delay in data transfer rate. Different research directions are discussed, which can lead toward basic changes in mobile networks design. Technologies like mmWave, MIMO, device-centric architectures can lead to form the basic pillars of 5G.

Yang et al. proposed the technique of heterogeneous network (HetNet), which creates a multinode topology in which many nodes of same attributes are deployed. In this perspective, the core components in HetNet are spatial modeling of nodes, associativity of cell users, and direct connection among devices. Figure 20.8 illustrates HetNet [2].

20.5.8 Multibeam-Based Communication System

MmWave has propagation limitations that are overcome by technologies like MIMO and beam-mmWave system. The basic structure for 5G is a multibeam-based communication system. Performance can be measured by the cell's capacity. Antenna array of massive MIMO and advanced beam technology is used to handle propagation limits of MMave band. 5G's basic structure for mobile network is known as a multibeam communication system. Advanced beam-forming technology has been introduced to evaluate cell capacity and disruption probability in mmWave systems with massive MIMO.

20.5.9 Software-Defined Networking (SDN)

Software defined networking (SDN) is one of the well-known techniques used by 5G. This technique partitions two functions of network: control and data-forwarding. Functions used by SDN are programmable and handle applications infrastructure, including services provided by a network. SDN is divided into three subgroups. The first is a software controller that is used to control network functions, such as a network's application programming interface (APIs), operating system, and global network view. The second group is southbound, and is dedicated to interface and protocols in the system's network. Third is the northbound group, which is responsible to provide an interface to applications of SDN and its controller.

20.6.1 Health Care

One of the most important and beneficial uses of 5G is in health care, by providing reliable and secure mobile communication for frequent data transmission from patient to cloud or to and from a remote health facility where the database for a specific patient has been maintained. A patient can get urgent and reliable medical service in the shortest possible span of time rather than by following the traditional procedures.

20.6.2 Smart Grid

5G can be pivotal in implementing smart grid technology, which focuses on better utilization of energy consumption by decentralizing energy distribution. Technologies used in 5G networks will help in smart distribution of power, providing efficient energy utilization.

20.6.3 Logistic and Tracking

5G is useful for location-based tracking. In future mobile communication systems, the tracking is easier. One of the practical examples of this application is radio frequency identification (RFID) tags.

20.6.4 Personal Usage

For heterogeneous networks, one of the key application of 5G is personal usage. It provides maximum use of UEs and meets high demand of not only text data but high-resolution video calls, voice calls, web surfing, and multimedia content, as well as providing QoS.

20.6.5 Virtualized Home

One of the most interesting applications of 5G is called virtualized home. They can be used by advances in wireless sensor networks. It will use C-RAN architecture with low-cost UEs with services of physical and data link layers.