Realizing 5G vision through Cloud RAN: technologies, challenges, and trends

Realizing 5G vision through Cloud RAN: technologies, challenges, and trends Achieving the fifth-generation (5G) vision will introduce new technology innovations and substantial changes in delivering cutting-edge applications and services in current mobile and cellular networks. The Cloud Radio Access Network (C-RAN) concept emerged as one of the most compelling architectures to meet the requirements of the 5G vision. In essence, C-RAN provides an advanced mobile network architecture which can leverage challenging features such as network resource slicing, statistical multiplexing, energy efficiency, and high capacity. The realization of C-RAN is achieved by innovative technologies such as the software-defined networking (SDN) and the network function virtualization (NFV). While SDN technology brings the separation of the control and data planes in the playground, supporting thus advanced traffic engineering techniques such as load balancing, the NFV concept offers high flexibility by allowing network resource sharing in a dynamic way. Although SDN and NFV have many advantages, a number of challenges have to be addressed before the commercial deployment of 5G implementation. In addition, C-RAN introduces a new layer in the mobile network, denoted as the fronthaul, which is adopted from the recent research efforts in the fiber-wireless (Fi-Wi) paradigm. As the fronthaul defines a link between a baseband unit (BBU) and a remote radio unit (RRU), various technologies can be used for this purpose such as optical fibers and millimeter-wave (mm-wave) radios. In this way, several challenges are highlighted which depend on the technology used. In the light of the aforementioned remarks, this paper compiles a list of challenges and open issues of the emerging technologies that realize the C-RAN concept. Moreover, comparative insights between the current and future state of the C-RAN concept are discussed. Trends and advances of those technologies are also examined towards shedding light on the proliferation of 5G through the C-RAN concept. Keywords: 5G, Cloud Radio Access Network, Common Public Radio Interface, Network function virtualization, Software-defined networking 1 Introduction very diverging requirements, energy-efficient communi- Mobile networks are rapidly evolving while the industry cations, and flexible and effective spectrum utilization. is struggling to keep up with the rising demand of con- Previous generations of mobile networks included stan- nectivity, data rates, capacity, and bandwidth. By 2021, it dard deployment schemes and fixed radio parameters is estimated that 10 billion devices will be connected to (e.g., frequency and power). However, 5G introduces sub- mobile networks worldwide [1], while the global data traf- stantial changes on many levels. It utilizes a broader spec- fic will rise to 49 EB per month. Future fifth-generation trum, using multiple access technologies, new deployment (5G) mobile networks are designed considering the fol- schemes (e.g., ultra-dense heterogeneous deployment), lowing fundamental requirements: massive connectivity, and advanced waveforms combined with novel coding and different traffic types, extremely high data rates, very modulation algorithms. high capacity, support for applications and services with The rapid proliferation of smart devices, along with the exponential rise in data traffic, creates a significant burden on current mobile networks. As current mobile network *Correspondence: psarigiannidis@uowm.gr 1 capacity is reaching its Shannon limit, operators try to sat- Department of Informatics and Telecommunications Engineering, University of Western Macedonia, Karamanli and Ligeris Str., GR-50100 Kozani, Greece isfy these requirements by deploying more base stations Full list of author information is available at the end of the article (BSs), creating thus a complex structure of ultra-dense © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 2 of 15 Table 1 List of acronyms heterogeneous networks. However, the mass deployment of ultra-dense BSs increases operators’ capital expendi- Acronym Definition ture (CAPEX)because of thecosts forsiteacquisition, 5G Fifth generation planning, and hardware equipment, as well as operating A/D Analog to digital expenditure (OPEX) due to maintenance and power usage API Application programming interface costs. As legacy Radio Access Network (RAN) is becom- BBU Baseband unit ing expensive and inadequate in satisfying the demands of future mobile trends, mobile operators are faced with the BS Base station challenge of devising new RAN architectures. CAPEX Capital expenditure Concepts such as cloud computing and virtualization CN Core network technologies are of paramount importance [2–4]since CoE CPRI over Ethernet they are deemed as candidate enablers for the new RAN CoMP Coordinated multipoint technologies. Cloud computing is a compelling concept CPRI Common public radio interface for enabling ubiquitous and on-demand access to a shared pool of scalable computing resources (e.g., network, stor- C-RAN Cloud radio access network age, and applications). Virtualization is realized through D/A Digital to analog two complementary technology concepts: network func- ETSI European telecommunications standards institute tion virtualization (NFV) [5, 6]and software-defined FFT Fast fourier transform networking (SDN) [7, 8]. SDN enables network pro- F-RAN Fog RAN gramming and provides network intelligence, while NFV HARQ Hybrid automatic repeat request leverages virtualization technologies to virtualize network functions. H-RAN Hybrid RAN Cloud-RAN (C-RAN) [9–11] is an innovative RAN IIoT Industrial Internet of Things technology based on the aforementioned concepts. In IoT Internet of Things C-RAN, operators can deploy mobile networks more ISG Industry specification group rapidly using different access technologies while sharing KPI Key performance indicator the same infrastructure. To this end, the deployment cost LTE Long-Term Evolution is reduced, the network resources are effectively utilized, and the maintenance cost is low. MIMO Multiple input-multiple output This work aims at presenting the key points of the mmWave Millimeter wave C-RAN architecture subject to the latest technologies and NFV Network function virtualization challenges. The C-RAN enabling technologies have been NFV MANO NFV management and orchestration extensively investigated in multiple studies. However, to NFVI NFV infrastructure the extent of our knowledge, this is the only work that OFDM Orthogonal frequency-division multiplexing compiles the challenges and open issues of all C-RAN components with respect to the 5G mobile networks. The ONOS Open network operating system rest of the paper is organized as follows. Table 1 lists all OPEX Operating expenditure the acronyms used throughout the article. Background ORI Open radio interface concepts related to the C-RAN implementation are pre- RAN Radio access network sented in Section 2.Section 3.1 provides an overview of RE Radio equipment C-RAN architecture and identifies its key components. REC Radio equipment controller Section 3.2 presents C-RAN state-of-art implementations. In Section 3.3, research challenges and open issues are RF Radio frequency discussed on C-RAN developments. Section 3.4 discusses RRU Remote radio head future trends and advancements. Finally, Section 4 con- SDMN Software-defined mobile network cludes this paper. SDN Software-defined networking T-SDN Transport SDN 2 Background VBBS Virtual big base station This section is devoted to presenting the background of the four main technologies behind C-RAN, namely VBS Virtual base station the SDN concept, the NFV technology, the network vir- VNF Virtualized network function tualization and slicing, and the Common Public Radio WDM Wavelength-division multiplexing Interface (CPRI) [12] which is the most widely used fron- WSN Wireless sensor network thaul interface in the C-RAN architecture. Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 3 of 15 2.1 Software-defined networking 2.2 Network function virtualization SDN has been primarily introduced for datacenter usage, Network function virtualization (NFV) is an emerging aiming to deliver flexibility in network deployment, oper- paradigm, which aims at offering new ways in designing, ation, and management [13]. The rationale behind SDN deploying, and managing modern network services [17]. is twofold: it separates the data plane from the control The main idea behind the NFV concept is to leverage vir- plane and it introduces novel network control functional- tualization technologies and decouple, thus, the network ity based on abstract network representation. The control functions from the physical equipment that accommo- decisions are removed from the hardware, and the net- dates them. This approach enables the concentration work intelligence is logically centralized. Network man- of network equipment and services in data centers, agement and operation are simplified through SDN, as where network functions run as software applications on forwarding and routing instructions are configured by general-purpose processor platforms. Furthermore, net- SDN controllers. work functions can be relocated at different network An overview of the SDN architecture is shown in locations without purchasing and installing new network Fig. 1. The application plane consists of the network equipment. applications (e.g., monitoring and security) and com- NFV architecture is shown in Fig. 2.Avirtualized municates with the control plane through the north- network function (VNF) is an implementation of a net- bound interface. The control plane consists of the work function (e.g., router and firewall) deployed on SDN controllers (e.g., Open Network Operating Sys- virtual resources provided by the NFV infrastructure tem (ONOS) [14]and OpenDayLight [15]) which gov- (NFVI). The NFVI contains the hardware and soft- ern the network devices. The devices are resided in the ware resources in which VNFs are deployed. It is com- data plane. The communication between control and posed by virtual and physical resources of storage, data planes is accomplished through the southbound computation, and network. The hypervisor is respon- interface. OpenFlow [16] is a widely adopted protocol sible for the mapping between the virtual and physi- for control and data plane communication. The data cal domains. The NFV management and orchestration plane consists of switches and routers which forward (NFV MANO) [18] framework is responsible for VNF the packets based on the configuration sent from the management and mapping between virtual and physical controllers. resources. Fig. 1 SDN architecture in a three-layer approach Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 4 of 15 Fig. 2 NFV architecture 2.3 Network virtualization and slicing key metrics are presented subject to the aforementioned Network virtualization allows operators to form their own scenarios, namely end-to-end latency, data rate, band- virtual networks. This is achieved through the concept of width, mobility, and number of connections per cell. Sen- network slicing, which allows network operators to create sor applications require a massive number of connections, end-to-end virtual networks which share the same phys- very low data rate and bandwidth, and very low mobility. ical infrastructure. A network slice is a virtual network On the contrary, vehicular applications require support created on top of a physical infrastructure in a way that for very frequent handovers. Industrial applications are the network operator believes that it performs on its own characterized by extremely low latency and high data rate dedicated physical network. and bandwidth. Smartphones are the most widely used The heterogeneity of modern service requirements are scenario with high overall requirements. Health applica- illustrated in Fig. 3. Five key characteristic application tions are an important category, containing critical appli- scenarios are highlighted, namely sensor, vehicular, indus- cations. They are characterized by low latency, medium trial, smartphone, and health paradigms. In addition, five bandwidth and data rate, and low mobility. Fig. 3 Application scenarios Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 5 of 15 Figure 3 highlights the necessity of incorporating virtualized and centralized into one entity called BBU network slicing, as virtual networks tailored to specific pool, which is often located at a data center, while RRHs requirements have better performance than typical are located at remote sites. A RRH is not attached to a multipurpose networks. Moreover, a network slice can single BBU and can be logically connected to any BBU be modified, depending on service requirements and from the BBU pool. The BBU pool consists of many BBUs number of users. Slices are also isolated to each other, which are deployed on servers with high-processing which enhances reliability and security, as configurations power. The BBUs operate as virtual base stations (VBSs), of different slices do not affect each other. which perform the baseband processing functions (e.g., fast Fourier transform (FFT)/inverse FFT, modula- 2.4 Common Public Radio Interface tion/demodulation, encoding/decoding, radio scheduling, CPRI is the standard interface that enables the commu- hybrid automatic repeat request (HARQ) management, nication between the radio equipment (RE) and the radio and radio link control). These functions are software equipment controller (REC) [19]. Point-to-point fiber is defined and they run as applications. Data from the BBU the most used physical transport technology for CPRI due pool are transported to the RRHs through a low-latency to its low cost and ease in installation. Data are transmit- and high-bandwidth interface called fronthaul. ted in the form of in-phase and quadrature signal (I/Q) RRHs transmit the RF signals to UEs, and they are flows, where each flow reflects the sampled and digitized responsible for radio frequency (RF) amplification, fil- radio signal of one carrier at one antenna element. The tering, and A/D and D/A conversion. As most of the standard specification defines hierarchical framing with processing functions are executed in the BBU pool, RRHs three layers so as to match the 3GPP Long-Term Evolu- are relatively simple and can be widely deployed in a tion (LTE) framing. The first layer is a CPRI basic frame, cost-efficient manner. which is transmitted every TC = 260.416 ns, based on In the initial C-RAN architecture, almost all baseband the 3.84-MHz clock rate. This basic frame consists of functionalities are moved to BBUs, while the RRHs act as 16 words, where word length depends on the CPRI con- a simple RF front-end. This split can achieve the highest figuration. The second layer is known as hyper-frame. processing gain but requires very high fronthaul band- It is a collection of 256 basic frames transmitted every width. Rather than offloading all baseband processing to 256 × TC = 66.67 ms, which is the LTE symbol time the BBU, it is possible to keep a subset of these functions [20], using orthogonal frequency-division multiplexing in the RRH [22]. The split can occur on any protocol layer. (OFDM) [21]. Finally, the third layer is a collection of 150 However, there are certain timing and capacity require- ments on inter-layer communication. The fronthaul link hyper-frames, which are created every 10 ms. The third layer carries the I/Q samples of a whole LTE frame. is a critical factor influencing the split level. Higher link quality and capacity allow a higher degree of centraliza- 3 Review tion, by moving more of the lower layer functions to the 3.1 Cloud-RAN cloud. This means that a trade-off between full central- C-RAN is based on the concepts of centralization and ization and fronthaul requirement satisfaction appears. virtualization. It intends to improve the overall network C-RAN has several advantages over traditional RAN as performance. It also reduces expenditures by leveraging described below: the network resources. Using cloud servers, operators can Advanced processing techniques: As BBUs are located scale up their deployments more rapidly, allowing differ- in powerful data centers, they have access to higher ent radio access technologies to share the same physical processing resources. Advanced processing techniques network infrastructure. The rest of this section provides can be easily implemented by leveraging these pro- an overview of the C-RAN architecture, its components, cessing resources. Coordinated multipoint (CoMP) [23] and the advantages over traditional RAN. processing is an effective technique to increase signal- In a C-RAN architecture, the LTE base station con- to-interference-plus-noise ratio (SINR), mitigate interfer- sists of the baseband unit (BBU) and the remote radio ence, and improve overall network throughput. head (RRH). The BBU performs baseband processing In [24], the authors aim to mitigate cell edge interfer- and provides higher layer functionality and communica- ence by adopting a clustered CoMP transmission scheme. tion with the core network. The RRH is responsible for Another approach to mitigate interference, by dividing radio functions, signal processing, modulation, analog-to- low- and high-mobility devices into clusters, was pre- digital (A/D) and digital-to-analog (D/A) conversion, and sented in [25]. The authors in [26]proposedanoptimiza- power amplification. tion framework for interference processing. In particular, the radio interference processing is formulated as short- The basic C-RAN architecture consists of three main term precoding and long-term user-centric RRH cluster- parts: the BBUs, the RRHs, and the fronthaul. The ing sub-problems. Hekrdla et al. [27]proposedanovel C-RAN architecture is shown in Fig. 4.BBUsare Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 6 of 15 Fig. 4 C-RAN architecture mechanism between multiple operators for downlink users. They consider the problem as a special case of mul- interference cancellation precoding. Inter-operator inter- tidimensional bin-packing problem, where each BBU is ference is canceled by adopting the regularized block viewed as bin and each virtual machine is viewed as item. diagonalization precoding to avoid user-sensitive data A similar approach was presented in [31]. The authors exchange between operators. Intra-operator interference solve the bin-packing problem using the best-fit decreas- is mitigated using Tomlinson-Harashima precoding with ing method, by jointly considering RRH resources and transmission power control. BBU scheduling. BBU scaling: BBUs are dynamically scaled according to Pompili et al. [32] proposed an elastic resource utiliza- the network requirements. For example, when there is an tion framework that aims to satisfy fluctuations in per- increase in network traffic, a virtual BBU can be scaled user capacity demands. They also introduced the idea of up to utilize more computing resources. In addition, in BBU clustering and discussed its advantages. The authors case of future network extensions, more virtual BBUs can in [33] proposed a scheme for inter-BBU load balancing be instantiated. A novel resource optimization algorithm in the BBU pool. The scheme involves a controller which which takes into account thermal and computing resource implements the inter-BBU management based on BBU models was developed in [28]. Optimization is achieved load threshold. by allocating the maximum load to BBU under thermal Energy efficiency: By having BBUs located at the data constraints. The optimization problem is solved using centers, the overall network energy consumption is Lagrange multiplier with Kuhn-Tucker condition. In [29], decreased. This is attached to the fact that the cell sites Zhang et al. aimed at minimizing the total amount of com- only include the RRHs, which have limited energy con- puting resources needed, while balancing the allocated sumption. Energy efficiency can also be increased by computing resources among BBUs. The optimization is dynamically managing (e.g., activation and operation) formulated as a bin-packing problem and solved using a BBUs depending on the data traffic demand and network heuristic genetic algorithm. load [34, 35]. Joint power control and user scheduling The authors in [30] studied the minimization of the techniques can highly augment energy efficiency as number of active BBUs that are required to serve the well [36]. Yu et al. [37] formulate the C-RAN energy Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 7 of 15 saving problem as a joint resource provisioning prob- efficiency, by having the RRHs negotiating with each other lem. Based on the traffic load, users and BBUs are reduce mutual interference. assigned to specific RRHs, so the energy consump- Big data analytics: Datacenter processing resources can tion of the entire system is minimized. The authors also be used for big data analytics. By analyzing mas- in [38] proposed a BBU-RRH assignment scheme that sive amounts of user and network data, operators can improves energy efficiency based on graph partitioning extract valuable findings regarding network performance and rejoining. Also, ultra-dense deployment reduces and quality of service. Proactive caching (e.g., popu- the distance between RRH and user, leading to high lar videos, images, and location-based content), based achievable data rates with low power consumption [39]. on user behavior prediction, significantly improves user The work in [40] presented a tunable distance-based experience and reduces network load [49]. Data collection power control mechanism, which improves energy can also be combined with machine learning techniques efficiency by reducing the energy consumed in the leading to a more intelligent, self-adaptive, and secure network nodes. network [50, 51]. An energy-efficient algorithm through the cloud-based An overview of the available mobile big data along with workload consolidation model was proposed in [41]. In a big data analytics-enabled network architecture was the proposed algorithm, workloads are distributed among proposed in [52]. Authors divide the data into four cate- virtualized BBUs that operate at full utilization, while gories, namely application, user, network, and link data. idle ones are turned off to reduce energy consumption. Application data describe features of applications such as The authors in [42] focused on network energy efficiency content popularity and service types. User data include through dynamic RRH activation and sparse beamform- user behavior, preferences, location, and mobility. Net- ing. In particular, they transform the energy efficiency work data contain configurations, signal strength, traffic maximization problem into a concave-convex functional load, and interference information. Finally, link data cover program based on weighted minimal mean square error physical channel information such as path loss, shadow- technique and group sparsity theory. ing, and channel statistics. Li et al. [43] designed a novel energy effective deploy- A similar classification approach was presented in ment scheme. The proposed scheme dynamically selects [53]. Data are classified into four categories, namely a subset of RRHs according to traffic demand and RRH flow record data, network performance data, mobile ter- capabilities. The RRH subset determination problem is minal data, and additional data. Flow record data are formulated as a multi-choice, multidimensional knapsack obtained through deep packet inspection and contain problem. The work in [44] formulated the RRH selec- the main attributes during a data session. Network per- tion problem as a trade-off between the minimization of formance data mainly include key performance indica- power consumption and transmission power, while satis- tor data and statistical information. They are used to fying a series of network constraints (i.e., spectrum limi- evaluate the network performance and the quality of tations and traffic requirements). An efficient local search service metrics. Mobile terminal data contain informa- algorithm was also proposed to address the formulated tion about the mobile devices, such as device informa- problem. tion, authentication information, cell identification, signal Resource scaling: In order to achieve optimal spectrum strength, and data rates. Additional data can be used efficiency, mobile devices should be attached to the BS at to build a subscriber profile for billing and data plan the best link quality [45]. C-RAN enables ultra-dense RRH information. deployment, providing mobile devices with more options The authors in [54]proposedanetworkoptimization for connection, while enabling reuse of time-frequency framework using big data analytics. They also discuss how resources. Provisioning of wireless resources is optimally the mobile big data are collected, stored, analyzed, and adapted to the actual needs of operators and subscribers. applied towards network optimization. In the same way, By exploiting centralized network intelligence, real-time Zhang et al. [55] discussed how big data analytics can resource allocation can be implemented in order to adapt be exploited to improve network performance in aspects to network conditions and user needs [46]. of network management, deployment, operation, and ser- Aiming to improve spectrum efficiency and decrease vice quality. By analyzing real-time network state informa- interference, the authors in [47]proposedajointcluster- tion, faults and anomalous behaviors can be predicted and ing and spectrum sharing scheme. They group virtualized mitigated. Spatial traffic load statistics can be obtained BSs into clusters and divide the frequency band into 13 from data analysis in order to determine the appropri- sub-bands. The sub-bands are allocated to mobile devices ate deployment of BSs. Real-time data traffic analysis can based on their position. extract traffic patterns in order to dynamically adjust the The authors in [48] proposed a novel approach based BSs resources, reducing energy consumption and improv- on a coalition formation game to improve the spectrum ing network operation. Finally, user quality of service can Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 8 of 15 be improved, by generating user behavior patterns based Most of the aforementioned efforts are focused on the on mobility and content data. RAN part of the mobile network. SoftRAN and FlexRAN Lower costs: Capital and operating expenditure can be handle interference and handover management aspects. reduced, as the BBUs are concentrated in data centers and FluidNet and FlexCRAN discuss the importance of a flex- the installation of RRHs requires less hardware. Moreover, ible fronthaul in the context of the C-RAN. SoftAir is the software and hardware upgrading becomes easier and less only work that is focused on both RAN and core network expensive [56]. (CN) parts of the mobile network. As a final note, Fluid- Net presents an IEEE 802.16-based implementation, while 3.2 C-RAN solutions FlexRAN and FlexCRAN present an OpenAirInterface This section lists, in chronological order, notable C-RAN LTE implementation. solutions that have been developed over the recent years. A discussion is also included at the end of this section. 3.3 Challenges and open issues SoftRAN [57] is considered as one of the earliest pro- In the previous sections, we presented an overview of posals in cloud and virtualization concept integration. All the C-RAN architecture and novel solutions. Still in its physical BSs are considered simple radio elements with infancy, it has many benefits but there are also several minimal control logic that form a virtual big BS (VBBS). challenges and open issues that need to be further investi- VBBS performs resource allocation, mobility, load balanc- gated in order to fully realize its potential. In the following ing, and other control functions. A logically centralized sections, we provide the challenges associated with each entity, i.e., the controller, maintains a global view of the of the C-RAN enabling technologies. Table 2 presents RAN and makes control plane decisions for all the RRHs. comparative insights between the current and the future In SoftAir [58], the control plane consists of network state of the C-RAN technology. In the following, a dis- management and optimization tools and it is imple- cussion is provided on the C-RAN challenges and the mented on the network servers. The data plane consists open issues. of software-defined BSs in the RAN and software-defined switches in the core network. Control and logic functions 3.3.1 Software-defined network are realized in software and executed on general-purpose The SDN concept is extended in order to support mobile processing platforms. The proposed architecture offers (i) communications. As software-defined mobile network programmability of the nodes, (ii) cooperation of nodes (SDMN) is a new notion and its specifications are still for enhancing network performance, (iii) open interface open, common SDN is used as a reference model for protocols, and (iv) an abstract view of the whole network SDMN design. There are critical issues that need to based on information collected from BSs and switches. be addressed to achieve seamless integration in the FluidNet [59] aims at providing intelligent configuration RAN [62]. of the fronthaul. FluidNet’s algorithms determine config- Architecture redesign: SDMN is different from con- urations that maximize the traffic demand satisfied on the ventional SDN, as mobile networks have fundamental RAN, while simultaneously optimizing the computation differences from wired networks. For example, the wire- resource usage in the BBU pool. less access domain is challenging because of the mas- FlexRAN [60] is a flexible and programmable software- sive number of devices and the heterogeneity of the defined RAN platform, which separates control from the modern mobile networks. Complex radio environments, data plane through a custom application programming which affect link reliability and quality, should also be interface (API). The main components are the FlexRAN taken into consideration while designing the SDMN master controller and the FlexRAN agent. Each agent cor- architecture. responds to a BS and is connected to the master controller. Controller placement: The controller placement heav- The FlexRAN API enables a two-way interaction between ily affects the network performance. Controller placement agents and the master controller. Agents act as local con- problem [63] aims at finding the optimal number of SDN trollers and send network state information to the master controllers as well as their location in order to minimize controller. Also, the master controller sends control com- the overhead latency and enhance the network reliabil- mands to the agents based on its knowledge of the entire ity. As stated before, SDMN architecture will introduce network state. additional constraints and requirements in the controller FlexCRAN [61] incorporates an architectural frame- placement problem. work that implements a flexible functional split, using Cognitive radio integration: Cognitive radio [64]isa Ethernet as fronthaul link. The authors also introduced promising approach in wireless communication engi- the key performance indicators (KPIs) of a C-RAN neering. Cognitive radio can monitor and dynamically and evaluated the proposed architecture through an reconfigure physical radio characteristics based on the OpenAirInterface-based implementation. environment variability. Centralized network intelligence Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 9 of 15 Table 2 Current and future state of C-RAN Concept Challenge Today Future Software-defined Architecture redesign Straightforward architecture using Complex architecture using wireless tech- networking wired technologies (i.e., Ethernet) for nologies. Radio environment affects link reli- interconnections. ability and quality. Controller placement Controllers are strategically placed in Wireless communications introduce addi- order to minimize latency. Increasing tional requirements in placement. Also, reli- the number of controllers also increases ability becomes more critical because of the control overhead. heterogeneity of the wireless environment. Cognitive radio integration Controllers maintain a high-level net- Network control is enhanced by adding work intelligence to achieve better net- radio environment intelligence to work control. controllers. Mobility management As the current SDN is used for conven- Mobility management is important, as users tional wired networks, there is no need should experience minimal disruptions in for mobility management. their communications. Network and function Performance optimization Hypervisors such as Kernel-based Virtual Utilize hypervisors optimized for extremely virtualization Machine (KVM) and Xen [99] are used for low overhead and latency. virtualizing resources. Network isolation Each virtual network configuration and Each virtual network configuration and cus- customization is independent from oth- tomization is independent from others. ers. Resource allocation Virtual machines access the physi- Spectrum availability is an additional cal resources through the hypervisor. resource feature that has to be managed. Computation, storage, and network Moreover, device mobility makes resource resources are the most common allocation more challenging. physical resources. Slice management Slices are scaled depending on service Slicing in 5G mobile networks is more chal- requirements. lenging as there are many operators sharing the same infrastructure and more diverging service requirements. Fronthaul Data reduction CPRI uses raw I/Q samples, requiring Techniques such as data compression, thus huge link capacity. aggregation, and redundancy removal should be considered. Latency reduction and Mobile communications require syn- Mobile communications require synchro- synchronization chronization and the lowest possible nization and lowest possible latency in order latency in order to ensure high quality of to ensure high quality of service. service. Overhead analysis CPRI has standard control signaling The trade-off between control overhead and information. bandwidth efficiency must be analyzed. It can also be dynamically adjusted depending on the link and network states. Novel standards CPRI is not the optimal fronthaul stan- Novel standards such as ORI should be dard for the C-RAN. researched. is an important benefit of conventional SDN. By inte- BBUs are shifted from physical machines to virtual grating cognitive radio, physical layer intelligence can be machines. As VMs use the same physical resources, obtained, resulting in better control of the overall net- the following challenges aim to minimize overhead and work. ensure high performance: Mobility management: Mobility management ensures Performance optimization: As network functions are successful data delivery and ensures no disruption in com- executed on top of a hypervisor (e.g., KVM and Xen), munications while the users are moving. Future mobile additional overhead is introduced, leading to network networks are expected to provide high-speed connectivity performance degradation. Optimization of legacy hyper- to extremely mobile applications (e.g., high-speed trains), visors and the introduction of new virtualization tech- complicating mobility management [65]. nologies (e.g., Docker) are plausible solutions towards minimizing overhead and latency. 3.3.2 Network and function virtualization Network isolation:Network isolationenables abstrac- tion and sharing of resources among different operators. Despite attracting wide attention by both industry and academia, the NFV concept still remains in early stages. Any configuration, customization, and topology change of Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 10 of 15 any virtual network should not affect and interfere other because of the high data rates required. Therefore, the coexisting operators. Isolation is challenging in wireless European Telecommunications Standards Institute (ETSI) networks due to the broadcast nature of wireless commu- has initiated a new Industry Specification Group (ISG) nications. For example, a change in a cell configuration called Open Radio Interface (ORI). The ORI goal is to may introduce high interference to neighbor cells. develop an interface specification envisioning interoper- Resource allocation: Resource allocation is another ability between parts of BSs of cellular mobile network significant challenge of wireless network virtualization. equipment. The interface defined by the ORI ISG is Unlike wired networks, resource allocation is more com- built on top of the CPRI with the removal and addi- plicated on wireless networks due to the availability tion of some options and functions in order to reach full of spectrum, device mobility, and the differentiation of interoperability. uplink and downlink channels. Slice management: Slice management and efficient allo- 3.4 Trends and advances cation of slice resources are interesting problems to be 3.4.1 SDN and NFV addressed [66]. Network slices should be created and Transport SDN [71]: The provisioning of infrastructure scaled dynamically based on service requirements. The to transport large data traffic between large geograph- resource allocation algorithms may adopt different strate- ical areas is too expensive for network operators. The gies depending on the slice size and service requirements. multitude advantages of SDN, i.e., cost reduction, urge The dynamic characteristics of 5G mobile networks must network operators to extend SDN utilization from data be taken into account while designing and implementing centers to large-scale geographic networks. SDN was orig- those algorithms. inally designed to operate at layers 2 and 3 of packet- switched networks. Transport SDN (T-SDN) extends the 3.3.3 Fronthaul SDN operation to the transport layer of circuit-switched Fronthaul has great impact on the performance of C- networks. T-SDN is in an initial stage with many chal- RAN [67]. To guarantee high network performance, links lenges and open issues. Due to its demand, T-SDN is with high data rate and bandwidth, as well as low latency gaining large attention from the industry and academia. and jitter, are required. Legacy fronthaul [68]has been Edge computing [72]: The proliferation of Internet of widely adopted in current network deployments, but it Things (IoT) devices introduces a substantial amount of will face serious challenges in the upcoming generation of data traffic in the mobile network. Instead of routing IoT mobile networks. The underlying focus of the following data to data centers for further processing, they are locally challenges is to achieve the lowest possible overhead and processed. As a result, the local process leads to network latency: traffic reduction. Edge computing is an emerging trend Data reduction: CPRI transports raw I/Q samples. An which aims at bringing the computational resources close 8 × 8 Multiple Input-Multiple Output (MIMO) system to the end devices. This shift of resources creates new with a 100-MHz channel (as it is envisioned in 5G) will complexities and challenges. require approximately 160 GB/s capacity. Data reduction Industrial Internet of Things: IoT is an indispensable techniques, such as data compression, aggregation, and segment of 5G mobile networks and a prerequisite for removal of redundancy should be implemented [69]. Industry 4.0 [73]. The Industrial IoT (IIoT) paradigm [74] Latency reduction and synchronization: Before being is envisioned to continuously obtain data from various processed, subframes need to be transported from RRHs sensors, transfer the data to cloud-based data centers for to BBUs. To ensure high quality of service, the time further processing, and update related configurations in required for the transportation should be extremely low. the industrial systems based on those data. This feedback- Synchronization information must be transmitted from based system will significantly raise the industry effi- BBUs to RRHs. Synchronization will be crucial in 5G ciency. As Wireless Sensor Networks (WSNs) are the basis mobile networks since access nodes are able to coop- of IIoT, research efforts are focused on the WSN virtu- erate with each other. For example, a frequency offset alization. A virtual WSN is formed by supporting logical between cooperating nodes will result in signal overlap, connectivity among collaborating sensors, based on either beamforming distortion, and degraded performance [70]. the information they track or the task they perform (e.g., Overhead analysis: Overhead should be quantified and traffic control and environmental monitoring). analyzed in order to identify design trade-offs. For exam- ple, smaller overhead may lead to reduced control infor- 3.4.2 Other transport technologies for fronthaul mation but increased bandwidth efficiency. Using point-to-point fiber, the number of links required Novel standards: CPRI is not an open standard while linearly increases with the number of RRHs deployed, it was originally designed as a BS internal interface. leading to capacity degradation, high cost, and increased Currently, it may not be the optimal interface, mainly network complexity. Authors in [75]discuss theurgency Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 11 of 15 of incorporating other transport technologies in order technology in next-generation mobile networks. Signal to address the aforementioned issues. They also evalu- propagation and channel characteristics have been exten- ate the capacity performance of fronthaul millimeter wave sively studied in the past. Nevertheless, [78]was oneof (mmWave) and provide a comparison of other potential the pioneer works that investigated mmWave’s feasibility fronthaul candidates. Table 3 provides a list of alterna- for next-generation mobile networks and demonstrated tive transport technologies along with the advantages and an early architecture. The authors in [79]conducted issues of each one. In the following, the alternative trans- extensive signal and channel measurements, ultimately port technologies are discussed in detail. concluding that mmWave will be a key communication Optical networks: Optical networks constitute a promis- technology in next-generation mobile networks. To that ing alternative solution for realizing the fronthaul of the end, many research studies (e.g., [80–82]) were devel- C-RAN since they provide high-capacity and low-latency oped in order to measure and compare mmWave’s per- links [76]. Moreover, they can accommodate multiple formance against conventional mobile technologies. The CPRI links in a single fiber by using techniques such authors in [83] carried out a survey of existing solu- as wavelength-division multiplexing (WDM), in which tions regarding standardization activities and discussed each link is associated with different wavelengths (chan- potential mmWave applications as well as open research nel). This reduces the number of fiber resources required. issues. An extensive survey of recent channel measure- Nonetheless, the network extensibility is still challenging ment and channel modeling results was presented in as it requires new fiber installation, if that is not already [84]. The authors also reviewed several technical chal- in place, while the most optical devices come with a lenges, namely severe pathloss and penetration loss, high cost. high power consumption, inconsistent main-lobe gain CPRI over Ethernet: The use of Ethernet for transport and non-zero side-lobe radiation, and hardware impair- in the fronthaul seems beneficial due to the maturity of ments. Rappaport et al. [85] emphasized the importance the technology and its wide adoption. Statistical mul- of developing accurate propagation models for long-term tiplexing and packet-based transport make CPRI over development of future mmWave systems and provided Ethernet (CoE) [77] efficient and cost effective. How- a comprehensive compilation of mmWave propagation ever, there are implementation challenges as packetization models. which introduces overhead and delay to the, already strict, mmWave bands hold much potential for use as wireless latency budget. Furthermore, Ethernet communication is fronthaul, mainly because of the high spectrum availabil- asynchronous, while fronthaul is characterized by strict ity [83]. The small wavelength enables placement of large synchronization requirements. arrays of antennas in a single transceiver, which encour- Millimeter-wave frequencies: Traditionally, mmWave ages the implementation of MIMO techniques, providing had been used for long-distance point-to-point commu- better SINR and improved spectrum efficiency. mmWave nications in satellite communications. Driven by its lat- also enables point-to-multipoint links [86], further reduc- est developments, mmWave is emerged as a promising ing the deployment cost. Rebato et al. [87] studied how the Table 3 Transport technologies for fronthaul Transport technology Advantages Issues Optical networks High-capacity Expensive optical devices Reduced fiber resources Requires previous infrastructure Passive (in the case of passive optical net- Challenging network extensibility working) CPRI-over-Ethernet Efficient Packetization introduces overhead and latency Cost-effective Ethernet is an asynchronous protocol, while fronthaul has strict synchronization requirements. Universal availability of interfaces Millimeter-wave frequencies High scalability and capacity High path loss High frequency reuse Low penetration Small form factor Line-of-sight requirement Underutilized spectrum Point-to-multipoint links Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 12 of 15 path loss affects the mobile network deployment in terms the authors propose a technique using edge caching. of small-cell density and transmission power. They also introduce an interference-aware price for edge In [88], the authors studied the path selection and the caching in order to effectively manage resource allocation. link scheduling problem for mmWave transport network Finally, they model the subchannel and power allocation as a mathematical optimization problem. Using mixed- problem as a non-cooperative game. integer linear programming, they proposed a joint opti- In Phantom Cell [97] architecture, small cells are over- mization for those objectives. laid on a macro cell. The macro cell uses lower frequencies The authors in [89] proposed a converged radio-over- and aims at ensuring wide coverage and high mobility, fiber solution and demonstrated two architectures. In while small cells use higher frequencies for providing high the first architecture, the BBus are connected via fiber data rates and capacity. with the mmWave RRHs, and those mmWave RRHs are The authors in [98] discuss on the load balancing, han- connected using mmWave links to small cells. Mobile dover, and interference issues that exist in conventional users are connected to those small cells using tradi- cell architectures. As cell size is shrinking, user movement tional frequency bands. In the second architecture, the lead to very frequent handovers which brings additional mmWave RRHs are directly connected to mobile users overhead and latency. A cell-less architecture was pre- using mmWave links. Moreover, Sung et al. [90]imple- sented, in order to cope with the aforementioned issues mented a fronthaul communication system based on and offer additional advantages such as superior traffic mmWave and assessed its performance. They also demon- management, improvement of coverage, and avoidance of strated a real-time operation of a 28-GHz mmWave-based frequent handovers. 5G prototype in order to confirm its technical feasibility. 4Conclusions 3.4.3 Advanced C-RAN architectures The upcoming arrival of 5G mobile networks introduces In Section 3.1, we described a primitive C-RAN architecture. extensive modifications to current RAN architectures. As C-RAN is still in the early stages of research and deploy- In this paper, we investigated the C-RAN concept, the ment, this architecture can be used as a precursor for more enabling technologies, and their key challenges. We also advanced architectures, as the ones presented below. presented C-RAN implementations and discussed future Hybrid-RAN (H-RAN) [91] is the combination of het- trends and advances. erogeneous network architecture with the C-RAN con- RAN is continuously evolving to provide higher data cept, aiming at further enhancing the performance of rates and capacity, efficient spectrum utilization, and mas- the mobile network. Heterogeneous network architec- sive and ubiquitous connectivity. Concepts such as SDN ture involves the deployment of small cells (e.g., micro-, and NFV as well as advancements in optical and wire- pico-, and femto-cells) and the utilization of multiple less technologies are shaping the way mobile networks are radio access technologies. This improves the network effi- designed and deployed, ultimately enabling operators to ciency and coverage by deploying small cells in areas with provide more diverse, flexible, and cost-effective services increased capacity demands or areas with bare coverage. to users. Aiming to offload C-RAN burden, the Fog Radio Access Acknowledgments Network (F-RAN) [92]isproposedasapossibleevolu- The authors thank the anonymous reviewers for their valuable comments and tion of C-RAN. Compared to C-RAN, collaborative radio suggestions. signal processing and cooperative radio resource manage- Authors’ contributions ment procedures in F-RANs are adaptively implemented All authors have equally contributed to the manuscript preparation. All at the edge devices, which are closer to the end users. authors read and approved the final manuscript. For example, by equipping RRHs with limited cache stor- Competing interests age, it is possible to pre-fetch popular content, avoiding The authors declare that they have no competing interests. thus congestion and delays during peak hours. This tech- Publisher’s Note nique is called edge caching [93]and is akey compo- Springer Nature remains neutral with regard to jurisdictional claims in nent in improving F-RAN’s performance. Authors in [94] published maps and institutional affiliations. present a F-RAN architecture using non-orthogonal mul- Author details tiple access (NOMA) [95], which can effectively leverage Department of Informatics and Telecommunications Engineering, University C-RAN’s advantages. They optimize the power alloca- of Western Macedonia, Karamanli and Ligeris Str., GR-50100 Kozani, Greece. Department of Physics, Aristotle University of Thessaloniki, GR-54124 tion problem under a non-cooperation framework and Thessaloniki, Greece. Department of Electrical and Computer Engineering, the subchannel allocation problem using a many-to-many Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece. two-side matching game algorithm. Zhang et al. [96]dis- Received: 17 January 2018 Accepted: 6 May 2018 cuss how the F-RAN attributes can assist in network load relief. For handover management and procedure, Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 13 of 15 References lte-advanced [coordinated and distributed mimo]. IEEE Wirel. Commun. 1. CV Forecast, Cisco Visual Networking Index: Global Mobile Data Traffic 17(3), 26–34 (2010) Forecast Update, 2016–2021 White Paper. Cisco Public Information (2017). 24. H Zhang, C Jiang, J Cheng, VC Leung, Cooperative interference mitigation https://www.cisco.com/c/en/us/solutions/collateral/service-provider/ and handover management for heterogeneous cloud small cell visual-networking-index-vni/mobile-white-paper-c11-520862.pdf networks. IEEE Wirel. Commun. 22(3), 92–99 (2015) 2. P Sarigiannidis, T Lagkas, S Bibi, A Ampatzoglou, P Bellavista, Hybrid 5g 25. A Hajisami, D Pompili, Dynamic joint processing: achieving high spectral optical-wireless sdn-based networks, challenges and open issues. IET efficiency in uplink 5g cellular networks. Comput. Netw. 126, 44–56 (2017) Netw. 6(6), 141–148 (2017) 26. A Liu, VK Lau, Two-timescale user-centric rrh clustering and precoding 3. P Rost, CJ Bernardos, A De Domenico, M Di Girolamo, M Lalam, A Maeder, optimization for cloud ran via local stochastic cutting plane. IEEE Trans. D Sabella, D Wübben, Cloud technologies for flexible 5g radio access Signal Process. 66(1), 64–76 (2018) networks. IEEE Commun. Mag. 52(5), 68–76 (2014) 27. M Hekrdla, A Matera, SHR Naqvi, U Spagnolini, in Communications 4. EJ Kitindi, S Fu, Y Jia, A Kabir, Y Wang, Wireless network virtualization with Workshops (ICC), 2016 IEEE International Conference On. sdn and c-ran for 5g networks: requirements, opportunities, and Interference-sharing multi-operator cooperation for downlink precoding challenges. IEEE Access. 5, 19099–19115 (2017) in cloud-ran architecture (IEEE, Kuala Lumpur, 2016), pp. 128–133 5. B Han, V Gopalakrishnan, L Ji, S Lee, Network function virtualization: 28. SKS Tyagi, T Lin, Y Zhou, in Vehicular Technology Conference (VTC Spring), challenges and opportunities for innovations. IEEE Commun. Mag. 53(2), 2017 IEEE 85th. Thermal-aware dynamic computing resource allocation for 90–97 (2015) bbu pool in centralized radio access networks (IEEE, Sydney, 2017), pp. 1–5 6. M Chiosi, D Clarke, P Willis, A Reid, J Feger, M Bugenhagen, W Khan, M 29. F Zhang, J Zheng, Y Zhang, L Chu, in Vehicular Technology Conference (VTC Fargano, C Cui, H Deng, et al., in SDN and OpenFlow SDN and OpenFlow Spring), 2017 IEEE 85th. An efficient and balanced bbu computing World Congress. Network functions virtualization, an introduction, resource allocation algorithm for cloud radio access networks (IEEE, benefits, enablers, challenges and call for action, (2012), pp. 22–24 Sydney, 2017), pp. 1–5 7. W Xia, Y Wen, CH Foh, D Niyato, H Xie, A survey on software-defined 30. NYu, ZSong, HDu,HHuang,XJia,in Communications (ICC), 2017 IEEE networking. IEEE Commun. Surv. Tutor. 17(1), 27–51 (2015) International Conference On. Multi-resource allocation in cloud radio 8. D Kreutz, FM Ramos, PE Verissimo, CE Rothenberg, S Azodolmolky, S access networks (IEEE, Paris, 2017), pp. 1–6 Uhlig, Software-defined networking: a comprehensive survey. Proc. IEEE. 31. K Wang, W Zhou, S Mao, On joint bbu/rrh resource allocation in 103(1), 14–76 (2015) heterogeneous cloud-rans. IEEE Internet Things J. 4(3), 749–759 (2017) 9. M Hadzialic, B Dosenovic, M Dzaferagic, J Musovic, in ELMAR, 2013 55th 32. D Pompili, A Hajisami, TX Tran, Elastic resource utilization framework for International Symposium. Cloud-ran: innovative radio access network high capacity and energy efficiency in cloud ran. IEEE Commun. Mag. architecture (IEEE, Zadar, 2013), pp. 115–120 54(1), 26–32 (2016) 10. J Wu, Z Zhang, Y Hong, Y Wen, Cloud radio access network (C-RAN): a 33. T Duan, M Zhang, Z Wang, C Song, in Computer and Communications primer. IEEE Netw. 29(1), 35–41 (2015) (ICCC), 2016 2nd IEEE International Conference On. Inter-bbu control 11. A Checko, HL Christiansen, Y Yan, L Scolari, G Kardaras, MS Berger, L mechanism for load balancing in c-ran-based bbu pool (IEEE, Chengdu, Dittmann, Cloud ran for mobile networks—a technology overview. IEEE 2016), pp. 2960–2964 Commun. Surv. Tutor. 17(1), 405–426 (2015) 34. H Guo, K Wang, H Ji, VC Leung, in Network Infrastructure and Digital 12. AB Ericsson, Huawei Technologies Co. Ltd, NEC Corporation, Alcatel Content (IC-NIDC), 2016 IEEE International Conference On.Energysavingin Lucent, Nokia Networks, Common public radio interface (cpri); interface c-ran based on bbu switching scheme (IEEE, Beijing, 2016), pp. 44–49 specification. CPRI Specification. 5, 1–119 (2015). http://www.cpri.info/ 35. C Bluemm, Y Zhang, P Alvarez, M Ruffini, LA DaSilva, in Communications downloads/CPRI_v_7_0_2015-10-09.pdf Workshops (ICC Workshops), 2017 IEEE International Conference On. 13. M Arslan, K Sundaresan, S Rangarajan, Software-defined networking in Dynamic energy savings in cloud-ran: an experimental assessment and cellular radio access networks: potential and challenges. IEEE Commun. implementation (IEEE, Paris, 2017), pp. 791–796 Mag. 53(1), 150–156 (2015) 36. S Samarakoon, M Bennis, W Saad, M Debbah, M Latva-aho, Ultra dense 14. PBerde,MGerola,JHart,YHiguchi,MKobayashi, TKoide,BLantz,B small cell networks: turning density into energy efficiency. IEEE J. Sel. O’Connor, P Radoslavov, W Snow, et al., in Proceedings of the Third Areas Commun. 34(5), 1267–1280 (2016) Workshop on Hot Topics in Software Defined Networking. Onos: towards an 37. N Yu, Z Song, H Du, H Huang, X Jia, Dynamic resource provisioning for open, distributed sdn os (ACM, Chicago, 2014), pp. 1–6 energy efficient cloud radio access networks. IEEE Trans. Cloud Comput. 15. J Medved, R Varga, A Tkacik, K Gray, in World of Wireless, Mobile and (2017). https://ieeexplore.ieee.org/document/7949048/ Multimedia Networks (WoWMoM), 2014 IEEE 15th International Symposium 38. BJ Sahu, S Dash, N Saxena, A Roy, Energy-efficient bbu allocation for on A. Opendaylight: towards a model-driven sdn controller architecture green c-ran. IEEE Commun. Lett. 21(7), 1637–1640 (2017) (IEEE, Sydney, 2014), pp. 1–6 39. D Feng, C Jiang, G Lim, LJ Cimini, G Feng, GY Li, A survey of 16. N McKeown, T Anderson, H Balakrishnan, G Parulkar, L Peterson, J Rexford, energy-efficient wireless communications. IEEE Commun. Surv. Tutor. S Shenker, J Turner, Openflow: enabling innovation in campus networks. 15(1), 167–178 (2013) ACM SIGCOMM Comput. Commun. Rev. 38(2), 69–74 (2008) 40. F Ghods, AO Fapojuwo, FM Ghannouchi, in Electrical and Computer 17. R Mijumbi, J Serrat, J-L Gorricho, N Bouten, F De Turck, R Boutaba, Engineering (CCECE), 2017 IEEE 30th Canadian Conference On.Energy Network function virtualization: state-of-the-art and research challenges. efficiency analysis of a c-ran with distance-based power control (IEEE, IEEE Commun. Surv. Tutor. 18(1), 236–262 (2016) Windsor, 2017), pp. 1–5 18. M Ersue, in Presentation at the IETF# 88 Meeting. Etsi nfv management and 41. T Sigwele, AS Alam, P Pillai, YF Hu, Energy-efficient cloud radio access orchestration—an overview, (Vancouver, 2013) networks by cloud based workload consolidation for 5g. J. Netw. Comput. 19. A de la Oliva, JA Hernandez, D Larrabeiti, A Azcorra, An overview of the Appl. 78, 1–8 (2017) CPRI specification and its application to C-RAN-based LTE scenarios. IEEE 42. Y Zeng, X Wen, Z Lu, Y Chen, H Shao, in Wireless Communication Systems Commun. Mag. 54(2), 152–159 (2016) (ISWCS), 2016 International Symposium On. Joint remote radio head 20. L ETSI, Evolved universal terrestrial radio access (E-UTRA); User equipment activation and beamforming for energy efficient c-ran (IEEE, Poznan, (UE) radio transmission and reception. ETSI TS. 136(101), 16 (2017) 2016), pp. 550–554 21. Y Wu, WY Zou, Orthogonal frequency division multiplexing: a multi-carrier 43. A Li, Y Sun, X Xu, C Yuan, in Computer Communications Workshops modulation scheme. IEEE Trans. Consum. Electron. 41(3), 392–399 (1995) (INFOCOM WKSHPS), 2016 IEEE Conference On. An energy-effective network 22. A Maeder, M Lalam, A De Domenico, E Pateromichelakis, D Wubben, J Bartelt, deployment scheme for 5g cloud radio access networks (IEEE, San R Fritzsche, P Rost, in 2014 European Conference on Networks and Francisco, 2016), pp. 684–689 Communications (EuCNC). Towards a flexible functional split for 44. W Zhao, S Wang, in Vehicular Technology Conference (VTC Spring), 2016 IEEE cloud-RAN networks (IEEE, Bologna, 2014), pp. 1–5 83rd. Remote radio head selection for power saving in cloud radio access 23. M Sawahashi, Y Kishiyama, A Morimoto, D Nishikawa, M Tanno, networks (IEEE, Nanjing, 2016), pp. 1–5 Coordinated multipoint transmission/reception techniques for Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 14 of 15 45. RQ Hu, Y Qian, An energy efficient and spectrum efficient wireless 69. S-H Park, O Simeone, O Sahin, SS Shitz, Fronthaul compression for cloud heterogeneous network framework for 5g systems. IEEE Commun. Mag. radio access networks: signal processing advances inspired by network 52(5), 94–101 (2014) information theory. IEEE Signal Proc. Mag. 31(6), 69–79 (2014) 46. S Parsaeefard, R Dawadi, M Derakhshani, T Le-Ngoc, M Baghani, Dynamic 70. A Checko, AC Juul, HL Christiansen, MS Berger, in Communication resource allocation for virtualized wireless networks in Workshop (ICCW), 2015 IEEE International Conference On. Synchronization massive-mimo-aided and fronthaul-limited c-ran. IEEE Trans. Veh. challenges in packet-based cloud-ran fronthaul for mobile networks (IEEE, Technol. 66(10), 9512–9520 (2017) London, 2015), pp. 2721–2726 47. A Hajisami, D Pompili, Joint virtual edge-clustering and spectrum 71. R Alvizu, G Maier, N Kukreja, A Pattavina, R Morro, A Capello, C Cavazzoni, allocation scheme for uplink interference mitigation in c-ran. Ad Hoc Comprehensive survey on T-SDN: software-defined networking for Netw. 72, 91–104 (2018) transport networks. IEEE Commun. Surv. Tutor. 19(4), 2232–2283 (2017) 48. Z Zhou, J Peng, X Zhang, K Liu, F Jiang, A game-theoretical approach for 72. AC Baktir, A Ozgovde, C Ersoy, How can edge computing benefit from spectrum efficiency improvement in cloud-ran. Mob. Inf. Syst. 2016,(2016). software-defined networking: a survey, use cases, and future directions. https://www.hindawi.com/journals/misy/2016/3068732/abs/ IEEE Commun. Surv. Tutor. 19(4), 2359–2391 (2017) 49. E Zeydan, E Bastug, M Bennis, MA Kader, IA Karatepe, AS Er, M Debbah, 73. M Wollschlaeger, T Sauter, J Jasperneite, The future of industrial Big data caching for networking: moving from cloud to edge. IEEE communication: automation networks in the era of the internet of things Commun. Mag. 54(9), 36–42 (2016) and industry 4.0. IEEE Ind. Electron. Mag. 11(1), 17–27 (2017) 50. A Imran, A Zoha, A Abu-Dayya, Challenges in 5g: how to empower son 74. MR Palattella, M Dohler, A Grieco, G Rizzo, J Torsner, T Engel, L Ladid, with big data for enabling 5g. IEEE Netw. 28(6), 27–33 (2014) Internet of Things in the 5G era: enablers, architecture, and business 51. P Sarigiannidis, A Sarigiannidis, I Moscholios, P Zwierzykowski, Diana: a models. IEEE J. Sel. Areas Commun. 34(3), 510–527 (2016) machine learning mechanism for adjusting the tdd uplink-downlink 75. H Zhang, Y Dong, J Cheng, MJ Hossain, VCM Leung, Fronthauling for 5g configuration in xg-pon-lte systems. Mob. Inf. Syst. 2017,(2017). https:// lte-u ultra dense cloud small cell networks. IEEE Wirel. Commun. 23(6), www.hindawi.com/journals/misy/2017/8198017/abs/ 48–53 (2016) 52. S Han, I Chih-Lin, G Li, S Wang, Q Sun, Big data enabled mobile network 76. N Cvijetic, in Telecommunications Network Strategy and Planning design for 5g and beyond. IEEE Commun. Mag. 55(9), 150–157 (2017) Symposium (Networks), 2014 16th International. Optical network evolution 53. X Zhang, Z Yi, Z Yan, G Min, W Wang, A Elmokashfi, S Maharjan, Y Zhang, for 5g mobile applications and sdn-based control (IEEE, Funchal, 2014), Social computing for mobile big data. Computer. 49(9), 86–90 (2016) pp. 1–5 54. K Zheng, Z Yang, K Zhang, P Chatzimisios, K Yang, W Xiang, Big 77. NJ Gomes, P Chanclou, P Turnbull, A Magee, V Jungnickel, Fronthaul data-driven optimization for mobile networks toward 5g. IEEE Netw. evolution: from cpri to ethernet. Opt. Fiber. Technol. 26, 50–58 (2015) 30(1), 44–51 (2016) 78. Z Pi, F Khan, An introduction to millimeter-wave mobile broadband 55. N Zhang, P Yang, J Ren, D Chen, L Yu, X Shen, Synergy of big data and 5g systems. IEEE Commun. Mag. 49(6), 101–107 (2011) wireless networks: opportunities, approaches, and challenges. IEEE Wirel. 79. TS Rappaport, S Sun, R Mayzus, H Zhao, Y Azar, K Wang, GN Wong, JK Commun. 25(1), 12–18 (2018) Schulz, M Samimi, F Gutierrez, Millimeter wave mobile communications 56. E Hernandez-Valencia, S Izzo, B Polonsky, How will nfv/sdn transform for 5g cellular: it will work!. IEEE Access. 1, 335–349 (2013) service provider opex? IEEE Netw. 29(3), 60–67 (2015) 80. S Rangan, TS Rappaport, E Erkip, Millimeter-wave cellular wireless 57. A Gudipati, D Perry, LE Li, S Katti, in Proceedings of the Second ACM networks: potentials and challenges. Proc. IEEE. 102(3), 366–385 (2014) SIGCOMM Workshop on Hot Topics in Software Defined Networking - HotSDN 81. W Hong, K-H Baek, Y Lee, Y Kim, S-T Ko, Study and prototyping of ’13. SoftRAN: software defined radio access network (ACM Press, New practically large-scale mmwave antenna systems for 5g cellular devices. York, USA, 2013), p. 25 IEEE Commun. Mag. 52(9), 63–69 (2014) 58. IF Akyildiz, P Wang, S-C Lin, SoftAir: a software defined networking 82. T Bai, RW Heath, Coverage and rate analysis for millimeter-wave cellular architecture for 5G wireless systems. Comput. Netw. 85, 1–18 (2015) networks. IEEE Trans. Wirel. Commun. 14(2), 1100–1114 (2015) 59. K Sundaresan, MY Arslan, S Singh, S Rangarajan, SV Krishnamurthy, 83. Y Niu, Y Li, D Jin, L Su, AV Vasilakos, A survey of millimeter wave FluidNet: a flexible cloud-based radio access network for small cells. communications (mmwave) for 5g: opportunities and challenges. Wirel. IEEE/ACM Trans. Netw. 24(2), 915–928 (2016) Netw. 21(8), 2657–2676 (2015) 60. X Foukas, N Nikaein, MM Kassem, MK Marina, K Kontovasilis, in Proceedings 84. M Xiao, S Mumtaz, Y Huang, L Dai, Y Li, M Matthaiou, GK Karagiannidis, of the 12th International on Conference on Emerging Networking E Björnson, K Yang, I Chih-Lin, et al, Millimeter wave communications for EXperiments and Technologies - CoNEXT ’16. FlexRAN: a flexible and future mobile networks. IEEE J. Sel. Areas Commun. 35(9), 1909–1935 programmable platform for software-defined radio access networks (2017) (ACM Press, New York, New York, USA, 2016), pp. 427–441 61. C-Y Chang, N Nikaein, R Knopp, T Spyropoulos, SS Kumar, in 2017 IEEE 85. TS Rappaport, Y Xing, GR MacCartney, AF Molisch, E Mellios, J Zhang, International Conference on Communications (ICC). FlexCRAN: a flexible Overview of millimeter wave communications for fifth-generation (5g) functional split framework over ethernet fronthaul in Cloud-RAN (IEEE, wireless networks—with a focus on propagation models. IEEE Trans. Paris, 2017), pp. 1–7 Antennas Propag. 65(12), 6213–6230 (2017) 62. K Pentikousis, Y Wang, W Hu, Mobileflow: toward software-defined 86. R Taori, A Sridharan, Point-to-multipoint in-band mmwave backhaul for mobile networks. IEEE Commun. Mag. 51(7), 44–53 (2013) 5g networks. IEEE Commun. Mag. 53(1), 195–201 (2015) 63. Bo, H, Youke, W, Chuan’an, W, Ying, W, in 2016 2nd IEEE International 87. M Rebato, M Mezzavilla, S Rangan, F Boccardi, M Zorzi, in European Conference on Computer and Communications (ICCC). The controller Wireless 2016; 22th European Wireless Conference; Proceedings Of. placement problem for software-defined networks (IEEE, Chengdu, 2016), Understanding noise and interference regimes in 5g millimeter-wave pp. 2435–2439 cellular networks (VDE, Oulu, 2016), pp. 1–5 64. Y-C Liang, K-C Chen, GY Li, P Mahonen, Cognitive radio networking and 88. D Huerfano, I Demirkol, P Legg, in Communications Workshops (ICC communications: an overview. IEEE Trans. Veh. Technol. 60(7), 3386–3407 Workshops), 2017 IEEE International Conference On. Joint optimization of (2011) path selection and link scheduling for millimeter wave transport 65. J Wu, P Fan, A survey on high mobility wireless communications: networks (IEEE, Paris, 2017), pp. 115–120 challenges, opportunities and solutions. IEEE Access. 4, 450–476 (2016) 89. S Papaioannou, G Kalfas, C Vagionas, C Mitsolidou, P Maniotis, A Miliou, 66. H Zhang, N Liu, X Chu, K Long, A-H Aghvami, VC Leung, Network slicing N Pleros, in Interactive Mobile Communication, Technologies and Learning. based 5g and future mobile networks: mobility, resource management, 5g small-cell networks exploiting optical technologies with mmwave and challenges. IEEE Commun. Mag. 55(8), 138–145 (2017) massive mimo and mt-mac protocols (Springer, 2017), pp. 805–813. 67. M Peng, C Wang, V Lau, HV Poor, Fronthaul-constrained cloud radio https://link.springer.com/chapter/10.1007/978-3-319-75175-7_79 access networks: insights and challenges. IEEE Wirel. Commun. 22(2), 90. M Sung, S-H Cho, J Kim, JK Lee, JH Lee, HS Chung, Demonstration of 152–160 (2015) 68. C-l I, Y Yuan, J Huang, S Ma, C Cui, R Duan, Rethink fronthaul for soft RAN. ifof-based mobile fronthaul in 5g prototype with 28-ghz millimeter wave. IEEE Commun. Mag. 53(9), 82–88 (2015) J. Light. Technol. 36(2), 601–609 (2018) Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 15 of 15 91. Peng, M, Li, Y, Jiang, J, Li, J, Wang, C, Heterogeneous cloud radio access networks: a new perspective for enhancing spectral and energy efficiencies. IEEE Wirel. Commun. 21(6), 126–135 (2014) 92. M Peng, S Yan, K Zhang, C Wang, Fog-computing-based radio access networks: issues and challenges. IEEE Netw. 30(4), 46–53 (2016) 93. D Liu, B Chen, C Yang, AF Molisch, Caching at the wireless edge: design aspects, challenges, and future directions. IEEE Commun. Mag. 54(9), 22–28 (2016) 94. H Zhang, Y Qiu, K Long, GK Karagiannidis, X Wang, A Nallanathan, Resource allocation in noma based fog radio access networks. IEEE Wirel. Commun (2018). arXiv preprint arXiv:1803.05641. https://arxiv.org/abs/ 1803.05641 95. Y Saito, Y Kishiyama, A Benjebbour, T Nakamura, A Li, K Higuchi, in Vehicular Technology Conference (VTC Spring), 2013 IEEE 77th. Non-orthogonal multiple access (noma) for cellular future radio access (IEEE, Dresden, 2013), pp. 1–5 96. H Zhang, Y Qiu, X Chu, K Long, VC Leung, Fog radio access networks: mobility management, interference mitigation, and resource optimization. IEEE Wirel. Commun. 24(6), 120–127 (2017) 97. T Asai, in 2015 International Conference on Optical Network Design and Modeling (ONDM). 5G radio access network and its requirements on mobile optical network (IEEE, Pisa, 2015), pp. 7–11 98. T Han, X Ge, L Wang, KS Kwak, Y Han, X Liu, 5G Converged cell-less communications in smart cities. IEEE Commun. Mag. 55(3), 44–50 (2017) 99. A Binu, GS Kumar, Virtualization techniques: a methodical review of xen and kvm. Adv. Comput. Commun. 190, 399–410 (2011). https://link. springer.com/chapter/10.1007/978-3-642-22709-7_40 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png EURASIP Journal on Wireless Communications and Networking Springer Journals

Realizing 5G vision through Cloud RAN: technologies, challenges, and trends

Free
15 pages
Loading next page...
 
/lp/springer_journal/realizing-5g-vision-through-cloud-ran-technologies-challenges-and-ahbwvbxxiy
Publisher
Springer International Publishing
Copyright
Copyright © 2018 by The Author(s)
Subject
Engineering; Signal,Image and Speech Processing; Communications Engineering, Networks; Information Systems Applications (incl.Internet)
eISSN
1687-1499
D.O.I.
10.1186/s13638-018-1142-1
Publisher site
See Article on Publisher Site

Abstract

Achieving the fifth-generation (5G) vision will introduce new technology innovations and substantial changes in delivering cutting-edge applications and services in current mobile and cellular networks. The Cloud Radio Access Network (C-RAN) concept emerged as one of the most compelling architectures to meet the requirements of the 5G vision. In essence, C-RAN provides an advanced mobile network architecture which can leverage challenging features such as network resource slicing, statistical multiplexing, energy efficiency, and high capacity. The realization of C-RAN is achieved by innovative technologies such as the software-defined networking (SDN) and the network function virtualization (NFV). While SDN technology brings the separation of the control and data planes in the playground, supporting thus advanced traffic engineering techniques such as load balancing, the NFV concept offers high flexibility by allowing network resource sharing in a dynamic way. Although SDN and NFV have many advantages, a number of challenges have to be addressed before the commercial deployment of 5G implementation. In addition, C-RAN introduces a new layer in the mobile network, denoted as the fronthaul, which is adopted from the recent research efforts in the fiber-wireless (Fi-Wi) paradigm. As the fronthaul defines a link between a baseband unit (BBU) and a remote radio unit (RRU), various technologies can be used for this purpose such as optical fibers and millimeter-wave (mm-wave) radios. In this way, several challenges are highlighted which depend on the technology used. In the light of the aforementioned remarks, this paper compiles a list of challenges and open issues of the emerging technologies that realize the C-RAN concept. Moreover, comparative insights between the current and future state of the C-RAN concept are discussed. Trends and advances of those technologies are also examined towards shedding light on the proliferation of 5G through the C-RAN concept. Keywords: 5G, Cloud Radio Access Network, Common Public Radio Interface, Network function virtualization, Software-defined networking 1 Introduction very diverging requirements, energy-efficient communi- Mobile networks are rapidly evolving while the industry cations, and flexible and effective spectrum utilization. is struggling to keep up with the rising demand of con- Previous generations of mobile networks included stan- nectivity, data rates, capacity, and bandwidth. By 2021, it dard deployment schemes and fixed radio parameters is estimated that 10 billion devices will be connected to (e.g., frequency and power). However, 5G introduces sub- mobile networks worldwide [1], while the global data traf- stantial changes on many levels. It utilizes a broader spec- fic will rise to 49 EB per month. Future fifth-generation trum, using multiple access technologies, new deployment (5G) mobile networks are designed considering the fol- schemes (e.g., ultra-dense heterogeneous deployment), lowing fundamental requirements: massive connectivity, and advanced waveforms combined with novel coding and different traffic types, extremely high data rates, very modulation algorithms. high capacity, support for applications and services with The rapid proliferation of smart devices, along with the exponential rise in data traffic, creates a significant burden on current mobile networks. As current mobile network *Correspondence: psarigiannidis@uowm.gr 1 capacity is reaching its Shannon limit, operators try to sat- Department of Informatics and Telecommunications Engineering, University of Western Macedonia, Karamanli and Ligeris Str., GR-50100 Kozani, Greece isfy these requirements by deploying more base stations Full list of author information is available at the end of the article (BSs), creating thus a complex structure of ultra-dense © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 2 of 15 Table 1 List of acronyms heterogeneous networks. However, the mass deployment of ultra-dense BSs increases operators’ capital expendi- Acronym Definition ture (CAPEX)because of thecosts forsiteacquisition, 5G Fifth generation planning, and hardware equipment, as well as operating A/D Analog to digital expenditure (OPEX) due to maintenance and power usage API Application programming interface costs. As legacy Radio Access Network (RAN) is becom- BBU Baseband unit ing expensive and inadequate in satisfying the demands of future mobile trends, mobile operators are faced with the BS Base station challenge of devising new RAN architectures. CAPEX Capital expenditure Concepts such as cloud computing and virtualization CN Core network technologies are of paramount importance [2–4]since CoE CPRI over Ethernet they are deemed as candidate enablers for the new RAN CoMP Coordinated multipoint technologies. Cloud computing is a compelling concept CPRI Common public radio interface for enabling ubiquitous and on-demand access to a shared pool of scalable computing resources (e.g., network, stor- C-RAN Cloud radio access network age, and applications). Virtualization is realized through D/A Digital to analog two complementary technology concepts: network func- ETSI European telecommunications standards institute tion virtualization (NFV) [5, 6]and software-defined FFT Fast fourier transform networking (SDN) [7, 8]. SDN enables network pro- F-RAN Fog RAN gramming and provides network intelligence, while NFV HARQ Hybrid automatic repeat request leverages virtualization technologies to virtualize network functions. H-RAN Hybrid RAN Cloud-RAN (C-RAN) [9–11] is an innovative RAN IIoT Industrial Internet of Things technology based on the aforementioned concepts. In IoT Internet of Things C-RAN, operators can deploy mobile networks more ISG Industry specification group rapidly using different access technologies while sharing KPI Key performance indicator the same infrastructure. To this end, the deployment cost LTE Long-Term Evolution is reduced, the network resources are effectively utilized, and the maintenance cost is low. MIMO Multiple input-multiple output This work aims at presenting the key points of the mmWave Millimeter wave C-RAN architecture subject to the latest technologies and NFV Network function virtualization challenges. The C-RAN enabling technologies have been NFV MANO NFV management and orchestration extensively investigated in multiple studies. However, to NFVI NFV infrastructure the extent of our knowledge, this is the only work that OFDM Orthogonal frequency-division multiplexing compiles the challenges and open issues of all C-RAN components with respect to the 5G mobile networks. The ONOS Open network operating system rest of the paper is organized as follows. Table 1 lists all OPEX Operating expenditure the acronyms used throughout the article. Background ORI Open radio interface concepts related to the C-RAN implementation are pre- RAN Radio access network sented in Section 2.Section 3.1 provides an overview of RE Radio equipment C-RAN architecture and identifies its key components. REC Radio equipment controller Section 3.2 presents C-RAN state-of-art implementations. In Section 3.3, research challenges and open issues are RF Radio frequency discussed on C-RAN developments. Section 3.4 discusses RRU Remote radio head future trends and advancements. Finally, Section 4 con- SDMN Software-defined mobile network cludes this paper. SDN Software-defined networking T-SDN Transport SDN 2 Background VBBS Virtual big base station This section is devoted to presenting the background of the four main technologies behind C-RAN, namely VBS Virtual base station the SDN concept, the NFV technology, the network vir- VNF Virtualized network function tualization and slicing, and the Common Public Radio WDM Wavelength-division multiplexing Interface (CPRI) [12] which is the most widely used fron- WSN Wireless sensor network thaul interface in the C-RAN architecture. Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 3 of 15 2.1 Software-defined networking 2.2 Network function virtualization SDN has been primarily introduced for datacenter usage, Network function virtualization (NFV) is an emerging aiming to deliver flexibility in network deployment, oper- paradigm, which aims at offering new ways in designing, ation, and management [13]. The rationale behind SDN deploying, and managing modern network services [17]. is twofold: it separates the data plane from the control The main idea behind the NFV concept is to leverage vir- plane and it introduces novel network control functional- tualization technologies and decouple, thus, the network ity based on abstract network representation. The control functions from the physical equipment that accommo- decisions are removed from the hardware, and the net- dates them. This approach enables the concentration work intelligence is logically centralized. Network man- of network equipment and services in data centers, agement and operation are simplified through SDN, as where network functions run as software applications on forwarding and routing instructions are configured by general-purpose processor platforms. Furthermore, net- SDN controllers. work functions can be relocated at different network An overview of the SDN architecture is shown in locations without purchasing and installing new network Fig. 1. The application plane consists of the network equipment. applications (e.g., monitoring and security) and com- NFV architecture is shown in Fig. 2.Avirtualized municates with the control plane through the north- network function (VNF) is an implementation of a net- bound interface. The control plane consists of the work function (e.g., router and firewall) deployed on SDN controllers (e.g., Open Network Operating Sys- virtual resources provided by the NFV infrastructure tem (ONOS) [14]and OpenDayLight [15]) which gov- (NFVI). The NFVI contains the hardware and soft- ern the network devices. The devices are resided in the ware resources in which VNFs are deployed. It is com- data plane. The communication between control and posed by virtual and physical resources of storage, data planes is accomplished through the southbound computation, and network. The hypervisor is respon- interface. OpenFlow [16] is a widely adopted protocol sible for the mapping between the virtual and physi- for control and data plane communication. The data cal domains. The NFV management and orchestration plane consists of switches and routers which forward (NFV MANO) [18] framework is responsible for VNF the packets based on the configuration sent from the management and mapping between virtual and physical controllers. resources. Fig. 1 SDN architecture in a three-layer approach Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 4 of 15 Fig. 2 NFV architecture 2.3 Network virtualization and slicing key metrics are presented subject to the aforementioned Network virtualization allows operators to form their own scenarios, namely end-to-end latency, data rate, band- virtual networks. This is achieved through the concept of width, mobility, and number of connections per cell. Sen- network slicing, which allows network operators to create sor applications require a massive number of connections, end-to-end virtual networks which share the same phys- very low data rate and bandwidth, and very low mobility. ical infrastructure. A network slice is a virtual network On the contrary, vehicular applications require support created on top of a physical infrastructure in a way that for very frequent handovers. Industrial applications are the network operator believes that it performs on its own characterized by extremely low latency and high data rate dedicated physical network. and bandwidth. Smartphones are the most widely used The heterogeneity of modern service requirements are scenario with high overall requirements. Health applica- illustrated in Fig. 3. Five key characteristic application tions are an important category, containing critical appli- scenarios are highlighted, namely sensor, vehicular, indus- cations. They are characterized by low latency, medium trial, smartphone, and health paradigms. In addition, five bandwidth and data rate, and low mobility. Fig. 3 Application scenarios Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 5 of 15 Figure 3 highlights the necessity of incorporating virtualized and centralized into one entity called BBU network slicing, as virtual networks tailored to specific pool, which is often located at a data center, while RRHs requirements have better performance than typical are located at remote sites. A RRH is not attached to a multipurpose networks. Moreover, a network slice can single BBU and can be logically connected to any BBU be modified, depending on service requirements and from the BBU pool. The BBU pool consists of many BBUs number of users. Slices are also isolated to each other, which are deployed on servers with high-processing which enhances reliability and security, as configurations power. The BBUs operate as virtual base stations (VBSs), of different slices do not affect each other. which perform the baseband processing functions (e.g., fast Fourier transform (FFT)/inverse FFT, modula- 2.4 Common Public Radio Interface tion/demodulation, encoding/decoding, radio scheduling, CPRI is the standard interface that enables the commu- hybrid automatic repeat request (HARQ) management, nication between the radio equipment (RE) and the radio and radio link control). These functions are software equipment controller (REC) [19]. Point-to-point fiber is defined and they run as applications. Data from the BBU the most used physical transport technology for CPRI due pool are transported to the RRHs through a low-latency to its low cost and ease in installation. Data are transmit- and high-bandwidth interface called fronthaul. ted in the form of in-phase and quadrature signal (I/Q) RRHs transmit the RF signals to UEs, and they are flows, where each flow reflects the sampled and digitized responsible for radio frequency (RF) amplification, fil- radio signal of one carrier at one antenna element. The tering, and A/D and D/A conversion. As most of the standard specification defines hierarchical framing with processing functions are executed in the BBU pool, RRHs three layers so as to match the 3GPP Long-Term Evolu- are relatively simple and can be widely deployed in a tion (LTE) framing. The first layer is a CPRI basic frame, cost-efficient manner. which is transmitted every TC = 260.416 ns, based on In the initial C-RAN architecture, almost all baseband the 3.84-MHz clock rate. This basic frame consists of functionalities are moved to BBUs, while the RRHs act as 16 words, where word length depends on the CPRI con- a simple RF front-end. This split can achieve the highest figuration. The second layer is known as hyper-frame. processing gain but requires very high fronthaul band- It is a collection of 256 basic frames transmitted every width. Rather than offloading all baseband processing to 256 × TC = 66.67 ms, which is the LTE symbol time the BBU, it is possible to keep a subset of these functions [20], using orthogonal frequency-division multiplexing in the RRH [22]. The split can occur on any protocol layer. (OFDM) [21]. Finally, the third layer is a collection of 150 However, there are certain timing and capacity require- ments on inter-layer communication. The fronthaul link hyper-frames, which are created every 10 ms. The third layer carries the I/Q samples of a whole LTE frame. is a critical factor influencing the split level. Higher link quality and capacity allow a higher degree of centraliza- 3 Review tion, by moving more of the lower layer functions to the 3.1 Cloud-RAN cloud. This means that a trade-off between full central- C-RAN is based on the concepts of centralization and ization and fronthaul requirement satisfaction appears. virtualization. It intends to improve the overall network C-RAN has several advantages over traditional RAN as performance. It also reduces expenditures by leveraging described below: the network resources. Using cloud servers, operators can Advanced processing techniques: As BBUs are located scale up their deployments more rapidly, allowing differ- in powerful data centers, they have access to higher ent radio access technologies to share the same physical processing resources. Advanced processing techniques network infrastructure. The rest of this section provides can be easily implemented by leveraging these pro- an overview of the C-RAN architecture, its components, cessing resources. Coordinated multipoint (CoMP) [23] and the advantages over traditional RAN. processing is an effective technique to increase signal- In a C-RAN architecture, the LTE base station con- to-interference-plus-noise ratio (SINR), mitigate interfer- sists of the baseband unit (BBU) and the remote radio ence, and improve overall network throughput. head (RRH). The BBU performs baseband processing In [24], the authors aim to mitigate cell edge interfer- and provides higher layer functionality and communica- ence by adopting a clustered CoMP transmission scheme. tion with the core network. The RRH is responsible for Another approach to mitigate interference, by dividing radio functions, signal processing, modulation, analog-to- low- and high-mobility devices into clusters, was pre- digital (A/D) and digital-to-analog (D/A) conversion, and sented in [25]. The authors in [26]proposedanoptimiza- power amplification. tion framework for interference processing. In particular, the radio interference processing is formulated as short- The basic C-RAN architecture consists of three main term precoding and long-term user-centric RRH cluster- parts: the BBUs, the RRHs, and the fronthaul. The ing sub-problems. Hekrdla et al. [27]proposedanovel C-RAN architecture is shown in Fig. 4.BBUsare Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 6 of 15 Fig. 4 C-RAN architecture mechanism between multiple operators for downlink users. They consider the problem as a special case of mul- interference cancellation precoding. Inter-operator inter- tidimensional bin-packing problem, where each BBU is ference is canceled by adopting the regularized block viewed as bin and each virtual machine is viewed as item. diagonalization precoding to avoid user-sensitive data A similar approach was presented in [31]. The authors exchange between operators. Intra-operator interference solve the bin-packing problem using the best-fit decreas- is mitigated using Tomlinson-Harashima precoding with ing method, by jointly considering RRH resources and transmission power control. BBU scheduling. BBU scaling: BBUs are dynamically scaled according to Pompili et al. [32] proposed an elastic resource utiliza- the network requirements. For example, when there is an tion framework that aims to satisfy fluctuations in per- increase in network traffic, a virtual BBU can be scaled user capacity demands. They also introduced the idea of up to utilize more computing resources. In addition, in BBU clustering and discussed its advantages. The authors case of future network extensions, more virtual BBUs can in [33] proposed a scheme for inter-BBU load balancing be instantiated. A novel resource optimization algorithm in the BBU pool. The scheme involves a controller which which takes into account thermal and computing resource implements the inter-BBU management based on BBU models was developed in [28]. Optimization is achieved load threshold. by allocating the maximum load to BBU under thermal Energy efficiency: By having BBUs located at the data constraints. The optimization problem is solved using centers, the overall network energy consumption is Lagrange multiplier with Kuhn-Tucker condition. In [29], decreased. This is attached to the fact that the cell sites Zhang et al. aimed at minimizing the total amount of com- only include the RRHs, which have limited energy con- puting resources needed, while balancing the allocated sumption. Energy efficiency can also be increased by computing resources among BBUs. The optimization is dynamically managing (e.g., activation and operation) formulated as a bin-packing problem and solved using a BBUs depending on the data traffic demand and network heuristic genetic algorithm. load [34, 35]. Joint power control and user scheduling The authors in [30] studied the minimization of the techniques can highly augment energy efficiency as number of active BBUs that are required to serve the well [36]. Yu et al. [37] formulate the C-RAN energy Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 7 of 15 saving problem as a joint resource provisioning prob- efficiency, by having the RRHs negotiating with each other lem. Based on the traffic load, users and BBUs are reduce mutual interference. assigned to specific RRHs, so the energy consump- Big data analytics: Datacenter processing resources can tion of the entire system is minimized. The authors also be used for big data analytics. By analyzing mas- in [38] proposed a BBU-RRH assignment scheme that sive amounts of user and network data, operators can improves energy efficiency based on graph partitioning extract valuable findings regarding network performance and rejoining. Also, ultra-dense deployment reduces and quality of service. Proactive caching (e.g., popu- the distance between RRH and user, leading to high lar videos, images, and location-based content), based achievable data rates with low power consumption [39]. on user behavior prediction, significantly improves user The work in [40] presented a tunable distance-based experience and reduces network load [49]. Data collection power control mechanism, which improves energy can also be combined with machine learning techniques efficiency by reducing the energy consumed in the leading to a more intelligent, self-adaptive, and secure network nodes. network [50, 51]. An energy-efficient algorithm through the cloud-based An overview of the available mobile big data along with workload consolidation model was proposed in [41]. In a big data analytics-enabled network architecture was the proposed algorithm, workloads are distributed among proposed in [52]. Authors divide the data into four cate- virtualized BBUs that operate at full utilization, while gories, namely application, user, network, and link data. idle ones are turned off to reduce energy consumption. Application data describe features of applications such as The authors in [42] focused on network energy efficiency content popularity and service types. User data include through dynamic RRH activation and sparse beamform- user behavior, preferences, location, and mobility. Net- ing. In particular, they transform the energy efficiency work data contain configurations, signal strength, traffic maximization problem into a concave-convex functional load, and interference information. Finally, link data cover program based on weighted minimal mean square error physical channel information such as path loss, shadow- technique and group sparsity theory. ing, and channel statistics. Li et al. [43] designed a novel energy effective deploy- A similar classification approach was presented in ment scheme. The proposed scheme dynamically selects [53]. Data are classified into four categories, namely a subset of RRHs according to traffic demand and RRH flow record data, network performance data, mobile ter- capabilities. The RRH subset determination problem is minal data, and additional data. Flow record data are formulated as a multi-choice, multidimensional knapsack obtained through deep packet inspection and contain problem. The work in [44] formulated the RRH selec- the main attributes during a data session. Network per- tion problem as a trade-off between the minimization of formance data mainly include key performance indica- power consumption and transmission power, while satis- tor data and statistical information. They are used to fying a series of network constraints (i.e., spectrum limi- evaluate the network performance and the quality of tations and traffic requirements). An efficient local search service metrics. Mobile terminal data contain informa- algorithm was also proposed to address the formulated tion about the mobile devices, such as device informa- problem. tion, authentication information, cell identification, signal Resource scaling: In order to achieve optimal spectrum strength, and data rates. Additional data can be used efficiency, mobile devices should be attached to the BS at to build a subscriber profile for billing and data plan the best link quality [45]. C-RAN enables ultra-dense RRH information. deployment, providing mobile devices with more options The authors in [54]proposedanetworkoptimization for connection, while enabling reuse of time-frequency framework using big data analytics. They also discuss how resources. Provisioning of wireless resources is optimally the mobile big data are collected, stored, analyzed, and adapted to the actual needs of operators and subscribers. applied towards network optimization. In the same way, By exploiting centralized network intelligence, real-time Zhang et al. [55] discussed how big data analytics can resource allocation can be implemented in order to adapt be exploited to improve network performance in aspects to network conditions and user needs [46]. of network management, deployment, operation, and ser- Aiming to improve spectrum efficiency and decrease vice quality. By analyzing real-time network state informa- interference, the authors in [47]proposedajointcluster- tion, faults and anomalous behaviors can be predicted and ing and spectrum sharing scheme. They group virtualized mitigated. Spatial traffic load statistics can be obtained BSs into clusters and divide the frequency band into 13 from data analysis in order to determine the appropri- sub-bands. The sub-bands are allocated to mobile devices ate deployment of BSs. Real-time data traffic analysis can based on their position. extract traffic patterns in order to dynamically adjust the The authors in [48] proposed a novel approach based BSs resources, reducing energy consumption and improv- on a coalition formation game to improve the spectrum ing network operation. Finally, user quality of service can Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 8 of 15 be improved, by generating user behavior patterns based Most of the aforementioned efforts are focused on the on mobility and content data. RAN part of the mobile network. SoftRAN and FlexRAN Lower costs: Capital and operating expenditure can be handle interference and handover management aspects. reduced, as the BBUs are concentrated in data centers and FluidNet and FlexCRAN discuss the importance of a flex- the installation of RRHs requires less hardware. Moreover, ible fronthaul in the context of the C-RAN. SoftAir is the software and hardware upgrading becomes easier and less only work that is focused on both RAN and core network expensive [56]. (CN) parts of the mobile network. As a final note, Fluid- Net presents an IEEE 802.16-based implementation, while 3.2 C-RAN solutions FlexRAN and FlexCRAN present an OpenAirInterface This section lists, in chronological order, notable C-RAN LTE implementation. solutions that have been developed over the recent years. A discussion is also included at the end of this section. 3.3 Challenges and open issues SoftRAN [57] is considered as one of the earliest pro- In the previous sections, we presented an overview of posals in cloud and virtualization concept integration. All the C-RAN architecture and novel solutions. Still in its physical BSs are considered simple radio elements with infancy, it has many benefits but there are also several minimal control logic that form a virtual big BS (VBBS). challenges and open issues that need to be further investi- VBBS performs resource allocation, mobility, load balanc- gated in order to fully realize its potential. In the following ing, and other control functions. A logically centralized sections, we provide the challenges associated with each entity, i.e., the controller, maintains a global view of the of the C-RAN enabling technologies. Table 2 presents RAN and makes control plane decisions for all the RRHs. comparative insights between the current and the future In SoftAir [58], the control plane consists of network state of the C-RAN technology. In the following, a dis- management and optimization tools and it is imple- cussion is provided on the C-RAN challenges and the mented on the network servers. The data plane consists open issues. of software-defined BSs in the RAN and software-defined switches in the core network. Control and logic functions 3.3.1 Software-defined network are realized in software and executed on general-purpose The SDN concept is extended in order to support mobile processing platforms. The proposed architecture offers (i) communications. As software-defined mobile network programmability of the nodes, (ii) cooperation of nodes (SDMN) is a new notion and its specifications are still for enhancing network performance, (iii) open interface open, common SDN is used as a reference model for protocols, and (iv) an abstract view of the whole network SDMN design. There are critical issues that need to based on information collected from BSs and switches. be addressed to achieve seamless integration in the FluidNet [59] aims at providing intelligent configuration RAN [62]. of the fronthaul. FluidNet’s algorithms determine config- Architecture redesign: SDMN is different from con- urations that maximize the traffic demand satisfied on the ventional SDN, as mobile networks have fundamental RAN, while simultaneously optimizing the computation differences from wired networks. For example, the wire- resource usage in the BBU pool. less access domain is challenging because of the mas- FlexRAN [60] is a flexible and programmable software- sive number of devices and the heterogeneity of the defined RAN platform, which separates control from the modern mobile networks. Complex radio environments, data plane through a custom application programming which affect link reliability and quality, should also be interface (API). The main components are the FlexRAN taken into consideration while designing the SDMN master controller and the FlexRAN agent. Each agent cor- architecture. responds to a BS and is connected to the master controller. Controller placement: The controller placement heav- The FlexRAN API enables a two-way interaction between ily affects the network performance. Controller placement agents and the master controller. Agents act as local con- problem [63] aims at finding the optimal number of SDN trollers and send network state information to the master controllers as well as their location in order to minimize controller. Also, the master controller sends control com- the overhead latency and enhance the network reliabil- mands to the agents based on its knowledge of the entire ity. As stated before, SDMN architecture will introduce network state. additional constraints and requirements in the controller FlexCRAN [61] incorporates an architectural frame- placement problem. work that implements a flexible functional split, using Cognitive radio integration: Cognitive radio [64]isa Ethernet as fronthaul link. The authors also introduced promising approach in wireless communication engi- the key performance indicators (KPIs) of a C-RAN neering. Cognitive radio can monitor and dynamically and evaluated the proposed architecture through an reconfigure physical radio characteristics based on the OpenAirInterface-based implementation. environment variability. Centralized network intelligence Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 9 of 15 Table 2 Current and future state of C-RAN Concept Challenge Today Future Software-defined Architecture redesign Straightforward architecture using Complex architecture using wireless tech- networking wired technologies (i.e., Ethernet) for nologies. Radio environment affects link reli- interconnections. ability and quality. Controller placement Controllers are strategically placed in Wireless communications introduce addi- order to minimize latency. Increasing tional requirements in placement. Also, reli- the number of controllers also increases ability becomes more critical because of the control overhead. heterogeneity of the wireless environment. Cognitive radio integration Controllers maintain a high-level net- Network control is enhanced by adding work intelligence to achieve better net- radio environment intelligence to work control. controllers. Mobility management As the current SDN is used for conven- Mobility management is important, as users tional wired networks, there is no need should experience minimal disruptions in for mobility management. their communications. Network and function Performance optimization Hypervisors such as Kernel-based Virtual Utilize hypervisors optimized for extremely virtualization Machine (KVM) and Xen [99] are used for low overhead and latency. virtualizing resources. Network isolation Each virtual network configuration and Each virtual network configuration and cus- customization is independent from oth- tomization is independent from others. ers. Resource allocation Virtual machines access the physi- Spectrum availability is an additional cal resources through the hypervisor. resource feature that has to be managed. Computation, storage, and network Moreover, device mobility makes resource resources are the most common allocation more challenging. physical resources. Slice management Slices are scaled depending on service Slicing in 5G mobile networks is more chal- requirements. lenging as there are many operators sharing the same infrastructure and more diverging service requirements. Fronthaul Data reduction CPRI uses raw I/Q samples, requiring Techniques such as data compression, thus huge link capacity. aggregation, and redundancy removal should be considered. Latency reduction and Mobile communications require syn- Mobile communications require synchro- synchronization chronization and the lowest possible nization and lowest possible latency in order latency in order to ensure high quality of to ensure high quality of service. service. Overhead analysis CPRI has standard control signaling The trade-off between control overhead and information. bandwidth efficiency must be analyzed. It can also be dynamically adjusted depending on the link and network states. Novel standards CPRI is not the optimal fronthaul stan- Novel standards such as ORI should be dard for the C-RAN. researched. is an important benefit of conventional SDN. By inte- BBUs are shifted from physical machines to virtual grating cognitive radio, physical layer intelligence can be machines. As VMs use the same physical resources, obtained, resulting in better control of the overall net- the following challenges aim to minimize overhead and work. ensure high performance: Mobility management: Mobility management ensures Performance optimization: As network functions are successful data delivery and ensures no disruption in com- executed on top of a hypervisor (e.g., KVM and Xen), munications while the users are moving. Future mobile additional overhead is introduced, leading to network networks are expected to provide high-speed connectivity performance degradation. Optimization of legacy hyper- to extremely mobile applications (e.g., high-speed trains), visors and the introduction of new virtualization tech- complicating mobility management [65]. nologies (e.g., Docker) are plausible solutions towards minimizing overhead and latency. 3.3.2 Network and function virtualization Network isolation:Network isolationenables abstrac- tion and sharing of resources among different operators. Despite attracting wide attention by both industry and academia, the NFV concept still remains in early stages. Any configuration, customization, and topology change of Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 10 of 15 any virtual network should not affect and interfere other because of the high data rates required. Therefore, the coexisting operators. Isolation is challenging in wireless European Telecommunications Standards Institute (ETSI) networks due to the broadcast nature of wireless commu- has initiated a new Industry Specification Group (ISG) nications. For example, a change in a cell configuration called Open Radio Interface (ORI). The ORI goal is to may introduce high interference to neighbor cells. develop an interface specification envisioning interoper- Resource allocation: Resource allocation is another ability between parts of BSs of cellular mobile network significant challenge of wireless network virtualization. equipment. The interface defined by the ORI ISG is Unlike wired networks, resource allocation is more com- built on top of the CPRI with the removal and addi- plicated on wireless networks due to the availability tion of some options and functions in order to reach full of spectrum, device mobility, and the differentiation of interoperability. uplink and downlink channels. Slice management: Slice management and efficient allo- 3.4 Trends and advances cation of slice resources are interesting problems to be 3.4.1 SDN and NFV addressed [66]. Network slices should be created and Transport SDN [71]: The provisioning of infrastructure scaled dynamically based on service requirements. The to transport large data traffic between large geograph- resource allocation algorithms may adopt different strate- ical areas is too expensive for network operators. The gies depending on the slice size and service requirements. multitude advantages of SDN, i.e., cost reduction, urge The dynamic characteristics of 5G mobile networks must network operators to extend SDN utilization from data be taken into account while designing and implementing centers to large-scale geographic networks. SDN was orig- those algorithms. inally designed to operate at layers 2 and 3 of packet- switched networks. Transport SDN (T-SDN) extends the 3.3.3 Fronthaul SDN operation to the transport layer of circuit-switched Fronthaul has great impact on the performance of C- networks. T-SDN is in an initial stage with many chal- RAN [67]. To guarantee high network performance, links lenges and open issues. Due to its demand, T-SDN is with high data rate and bandwidth, as well as low latency gaining large attention from the industry and academia. and jitter, are required. Legacy fronthaul [68]has been Edge computing [72]: The proliferation of Internet of widely adopted in current network deployments, but it Things (IoT) devices introduces a substantial amount of will face serious challenges in the upcoming generation of data traffic in the mobile network. Instead of routing IoT mobile networks. The underlying focus of the following data to data centers for further processing, they are locally challenges is to achieve the lowest possible overhead and processed. As a result, the local process leads to network latency: traffic reduction. Edge computing is an emerging trend Data reduction: CPRI transports raw I/Q samples. An which aims at bringing the computational resources close 8 × 8 Multiple Input-Multiple Output (MIMO) system to the end devices. This shift of resources creates new with a 100-MHz channel (as it is envisioned in 5G) will complexities and challenges. require approximately 160 GB/s capacity. Data reduction Industrial Internet of Things: IoT is an indispensable techniques, such as data compression, aggregation, and segment of 5G mobile networks and a prerequisite for removal of redundancy should be implemented [69]. Industry 4.0 [73]. The Industrial IoT (IIoT) paradigm [74] Latency reduction and synchronization: Before being is envisioned to continuously obtain data from various processed, subframes need to be transported from RRHs sensors, transfer the data to cloud-based data centers for to BBUs. To ensure high quality of service, the time further processing, and update related configurations in required for the transportation should be extremely low. the industrial systems based on those data. This feedback- Synchronization information must be transmitted from based system will significantly raise the industry effi- BBUs to RRHs. Synchronization will be crucial in 5G ciency. As Wireless Sensor Networks (WSNs) are the basis mobile networks since access nodes are able to coop- of IIoT, research efforts are focused on the WSN virtu- erate with each other. For example, a frequency offset alization. A virtual WSN is formed by supporting logical between cooperating nodes will result in signal overlap, connectivity among collaborating sensors, based on either beamforming distortion, and degraded performance [70]. the information they track or the task they perform (e.g., Overhead analysis: Overhead should be quantified and traffic control and environmental monitoring). analyzed in order to identify design trade-offs. For exam- ple, smaller overhead may lead to reduced control infor- 3.4.2 Other transport technologies for fronthaul mation but increased bandwidth efficiency. Using point-to-point fiber, the number of links required Novel standards: CPRI is not an open standard while linearly increases with the number of RRHs deployed, it was originally designed as a BS internal interface. leading to capacity degradation, high cost, and increased Currently, it may not be the optimal interface, mainly network complexity. Authors in [75]discuss theurgency Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 11 of 15 of incorporating other transport technologies in order technology in next-generation mobile networks. Signal to address the aforementioned issues. They also evalu- propagation and channel characteristics have been exten- ate the capacity performance of fronthaul millimeter wave sively studied in the past. Nevertheless, [78]was oneof (mmWave) and provide a comparison of other potential the pioneer works that investigated mmWave’s feasibility fronthaul candidates. Table 3 provides a list of alterna- for next-generation mobile networks and demonstrated tive transport technologies along with the advantages and an early architecture. The authors in [79]conducted issues of each one. In the following, the alternative trans- extensive signal and channel measurements, ultimately port technologies are discussed in detail. concluding that mmWave will be a key communication Optical networks: Optical networks constitute a promis- technology in next-generation mobile networks. To that ing alternative solution for realizing the fronthaul of the end, many research studies (e.g., [80–82]) were devel- C-RAN since they provide high-capacity and low-latency oped in order to measure and compare mmWave’s per- links [76]. Moreover, they can accommodate multiple formance against conventional mobile technologies. The CPRI links in a single fiber by using techniques such authors in [83] carried out a survey of existing solu- as wavelength-division multiplexing (WDM), in which tions regarding standardization activities and discussed each link is associated with different wavelengths (chan- potential mmWave applications as well as open research nel). This reduces the number of fiber resources required. issues. An extensive survey of recent channel measure- Nonetheless, the network extensibility is still challenging ment and channel modeling results was presented in as it requires new fiber installation, if that is not already [84]. The authors also reviewed several technical chal- in place, while the most optical devices come with a lenges, namely severe pathloss and penetration loss, high cost. high power consumption, inconsistent main-lobe gain CPRI over Ethernet: The use of Ethernet for transport and non-zero side-lobe radiation, and hardware impair- in the fronthaul seems beneficial due to the maturity of ments. Rappaport et al. [85] emphasized the importance the technology and its wide adoption. Statistical mul- of developing accurate propagation models for long-term tiplexing and packet-based transport make CPRI over development of future mmWave systems and provided Ethernet (CoE) [77] efficient and cost effective. How- a comprehensive compilation of mmWave propagation ever, there are implementation challenges as packetization models. which introduces overhead and delay to the, already strict, mmWave bands hold much potential for use as wireless latency budget. Furthermore, Ethernet communication is fronthaul, mainly because of the high spectrum availabil- asynchronous, while fronthaul is characterized by strict ity [83]. The small wavelength enables placement of large synchronization requirements. arrays of antennas in a single transceiver, which encour- Millimeter-wave frequencies: Traditionally, mmWave ages the implementation of MIMO techniques, providing had been used for long-distance point-to-point commu- better SINR and improved spectrum efficiency. mmWave nications in satellite communications. Driven by its lat- also enables point-to-multipoint links [86], further reduc- est developments, mmWave is emerged as a promising ing the deployment cost. Rebato et al. [87] studied how the Table 3 Transport technologies for fronthaul Transport technology Advantages Issues Optical networks High-capacity Expensive optical devices Reduced fiber resources Requires previous infrastructure Passive (in the case of passive optical net- Challenging network extensibility working) CPRI-over-Ethernet Efficient Packetization introduces overhead and latency Cost-effective Ethernet is an asynchronous protocol, while fronthaul has strict synchronization requirements. Universal availability of interfaces Millimeter-wave frequencies High scalability and capacity High path loss High frequency reuse Low penetration Small form factor Line-of-sight requirement Underutilized spectrum Point-to-multipoint links Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 12 of 15 path loss affects the mobile network deployment in terms the authors propose a technique using edge caching. of small-cell density and transmission power. They also introduce an interference-aware price for edge In [88], the authors studied the path selection and the caching in order to effectively manage resource allocation. link scheduling problem for mmWave transport network Finally, they model the subchannel and power allocation as a mathematical optimization problem. Using mixed- problem as a non-cooperative game. integer linear programming, they proposed a joint opti- In Phantom Cell [97] architecture, small cells are over- mization for those objectives. laid on a macro cell. The macro cell uses lower frequencies The authors in [89] proposed a converged radio-over- and aims at ensuring wide coverage and high mobility, fiber solution and demonstrated two architectures. In while small cells use higher frequencies for providing high the first architecture, the BBus are connected via fiber data rates and capacity. with the mmWave RRHs, and those mmWave RRHs are The authors in [98] discuss on the load balancing, han- connected using mmWave links to small cells. Mobile dover, and interference issues that exist in conventional users are connected to those small cells using tradi- cell architectures. As cell size is shrinking, user movement tional frequency bands. In the second architecture, the lead to very frequent handovers which brings additional mmWave RRHs are directly connected to mobile users overhead and latency. A cell-less architecture was pre- using mmWave links. Moreover, Sung et al. [90]imple- sented, in order to cope with the aforementioned issues mented a fronthaul communication system based on and offer additional advantages such as superior traffic mmWave and assessed its performance. They also demon- management, improvement of coverage, and avoidance of strated a real-time operation of a 28-GHz mmWave-based frequent handovers. 5G prototype in order to confirm its technical feasibility. 4Conclusions 3.4.3 Advanced C-RAN architectures The upcoming arrival of 5G mobile networks introduces In Section 3.1, we described a primitive C-RAN architecture. extensive modifications to current RAN architectures. As C-RAN is still in the early stages of research and deploy- In this paper, we investigated the C-RAN concept, the ment, this architecture can be used as a precursor for more enabling technologies, and their key challenges. We also advanced architectures, as the ones presented below. presented C-RAN implementations and discussed future Hybrid-RAN (H-RAN) [91] is the combination of het- trends and advances. erogeneous network architecture with the C-RAN con- RAN is continuously evolving to provide higher data cept, aiming at further enhancing the performance of rates and capacity, efficient spectrum utilization, and mas- the mobile network. Heterogeneous network architec- sive and ubiquitous connectivity. Concepts such as SDN ture involves the deployment of small cells (e.g., micro-, and NFV as well as advancements in optical and wire- pico-, and femto-cells) and the utilization of multiple less technologies are shaping the way mobile networks are radio access technologies. This improves the network effi- designed and deployed, ultimately enabling operators to ciency and coverage by deploying small cells in areas with provide more diverse, flexible, and cost-effective services increased capacity demands or areas with bare coverage. to users. Aiming to offload C-RAN burden, the Fog Radio Access Acknowledgments Network (F-RAN) [92]isproposedasapossibleevolu- The authors thank the anonymous reviewers for their valuable comments and tion of C-RAN. Compared to C-RAN, collaborative radio suggestions. signal processing and cooperative radio resource manage- Authors’ contributions ment procedures in F-RANs are adaptively implemented All authors have equally contributed to the manuscript preparation. All at the edge devices, which are closer to the end users. authors read and approved the final manuscript. For example, by equipping RRHs with limited cache stor- Competing interests age, it is possible to pre-fetch popular content, avoiding The authors declare that they have no competing interests. thus congestion and delays during peak hours. This tech- Publisher’s Note nique is called edge caching [93]and is akey compo- Springer Nature remains neutral with regard to jurisdictional claims in nent in improving F-RAN’s performance. Authors in [94] published maps and institutional affiliations. present a F-RAN architecture using non-orthogonal mul- Author details tiple access (NOMA) [95], which can effectively leverage Department of Informatics and Telecommunications Engineering, University C-RAN’s advantages. They optimize the power alloca- of Western Macedonia, Karamanli and Ligeris Str., GR-50100 Kozani, Greece. Department of Physics, Aristotle University of Thessaloniki, GR-54124 tion problem under a non-cooperation framework and Thessaloniki, Greece. Department of Electrical and Computer Engineering, the subchannel allocation problem using a many-to-many Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece. two-side matching game algorithm. Zhang et al. [96]dis- Received: 17 January 2018 Accepted: 6 May 2018 cuss how the F-RAN attributes can assist in network load relief. For handover management and procedure, Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 13 of 15 References lte-advanced [coordinated and distributed mimo]. IEEE Wirel. Commun. 1. CV Forecast, Cisco Visual Networking Index: Global Mobile Data Traffic 17(3), 26–34 (2010) Forecast Update, 2016–2021 White Paper. Cisco Public Information (2017). 24. H Zhang, C Jiang, J Cheng, VC Leung, Cooperative interference mitigation https://www.cisco.com/c/en/us/solutions/collateral/service-provider/ and handover management for heterogeneous cloud small cell visual-networking-index-vni/mobile-white-paper-c11-520862.pdf networks. IEEE Wirel. Commun. 22(3), 92–99 (2015) 2. P Sarigiannidis, T Lagkas, S Bibi, A Ampatzoglou, P Bellavista, Hybrid 5g 25. A Hajisami, D Pompili, Dynamic joint processing: achieving high spectral optical-wireless sdn-based networks, challenges and open issues. IET efficiency in uplink 5g cellular networks. Comput. Netw. 126, 44–56 (2017) Netw. 6(6), 141–148 (2017) 26. A Liu, VK Lau, Two-timescale user-centric rrh clustering and precoding 3. P Rost, CJ Bernardos, A De Domenico, M Di Girolamo, M Lalam, A Maeder, optimization for cloud ran via local stochastic cutting plane. IEEE Trans. D Sabella, D Wübben, Cloud technologies for flexible 5g radio access Signal Process. 66(1), 64–76 (2018) networks. IEEE Commun. Mag. 52(5), 68–76 (2014) 27. M Hekrdla, A Matera, SHR Naqvi, U Spagnolini, in Communications 4. EJ Kitindi, S Fu, Y Jia, A Kabir, Y Wang, Wireless network virtualization with Workshops (ICC), 2016 IEEE International Conference On. sdn and c-ran for 5g networks: requirements, opportunities, and Interference-sharing multi-operator cooperation for downlink precoding challenges. IEEE Access. 5, 19099–19115 (2017) in cloud-ran architecture (IEEE, Kuala Lumpur, 2016), pp. 128–133 5. B Han, V Gopalakrishnan, L Ji, S Lee, Network function virtualization: 28. SKS Tyagi, T Lin, Y Zhou, in Vehicular Technology Conference (VTC Spring), challenges and opportunities for innovations. IEEE Commun. Mag. 53(2), 2017 IEEE 85th. Thermal-aware dynamic computing resource allocation for 90–97 (2015) bbu pool in centralized radio access networks (IEEE, Sydney, 2017), pp. 1–5 6. M Chiosi, D Clarke, P Willis, A Reid, J Feger, M Bugenhagen, W Khan, M 29. F Zhang, J Zheng, Y Zhang, L Chu, in Vehicular Technology Conference (VTC Fargano, C Cui, H Deng, et al., in SDN and OpenFlow SDN and OpenFlow Spring), 2017 IEEE 85th. An efficient and balanced bbu computing World Congress. Network functions virtualization, an introduction, resource allocation algorithm for cloud radio access networks (IEEE, benefits, enablers, challenges and call for action, (2012), pp. 22–24 Sydney, 2017), pp. 1–5 7. W Xia, Y Wen, CH Foh, D Niyato, H Xie, A survey on software-defined 30. NYu, ZSong, HDu,HHuang,XJia,in Communications (ICC), 2017 IEEE networking. IEEE Commun. Surv. Tutor. 17(1), 27–51 (2015) International Conference On. Multi-resource allocation in cloud radio 8. D Kreutz, FM Ramos, PE Verissimo, CE Rothenberg, S Azodolmolky, S access networks (IEEE, Paris, 2017), pp. 1–6 Uhlig, Software-defined networking: a comprehensive survey. Proc. IEEE. 31. K Wang, W Zhou, S Mao, On joint bbu/rrh resource allocation in 103(1), 14–76 (2015) heterogeneous cloud-rans. IEEE Internet Things J. 4(3), 749–759 (2017) 9. M Hadzialic, B Dosenovic, M Dzaferagic, J Musovic, in ELMAR, 2013 55th 32. D Pompili, A Hajisami, TX Tran, Elastic resource utilization framework for International Symposium. Cloud-ran: innovative radio access network high capacity and energy efficiency in cloud ran. IEEE Commun. Mag. architecture (IEEE, Zadar, 2013), pp. 115–120 54(1), 26–32 (2016) 10. J Wu, Z Zhang, Y Hong, Y Wen, Cloud radio access network (C-RAN): a 33. T Duan, M Zhang, Z Wang, C Song, in Computer and Communications primer. IEEE Netw. 29(1), 35–41 (2015) (ICCC), 2016 2nd IEEE International Conference On. Inter-bbu control 11. A Checko, HL Christiansen, Y Yan, L Scolari, G Kardaras, MS Berger, L mechanism for load balancing in c-ran-based bbu pool (IEEE, Chengdu, Dittmann, Cloud ran for mobile networks—a technology overview. IEEE 2016), pp. 2960–2964 Commun. Surv. Tutor. 17(1), 405–426 (2015) 34. H Guo, K Wang, H Ji, VC Leung, in Network Infrastructure and Digital 12. AB Ericsson, Huawei Technologies Co. Ltd, NEC Corporation, Alcatel Content (IC-NIDC), 2016 IEEE International Conference On.Energysavingin Lucent, Nokia Networks, Common public radio interface (cpri); interface c-ran based on bbu switching scheme (IEEE, Beijing, 2016), pp. 44–49 specification. CPRI Specification. 5, 1–119 (2015). http://www.cpri.info/ 35. C Bluemm, Y Zhang, P Alvarez, M Ruffini, LA DaSilva, in Communications downloads/CPRI_v_7_0_2015-10-09.pdf Workshops (ICC Workshops), 2017 IEEE International Conference On. 13. M Arslan, K Sundaresan, S Rangarajan, Software-defined networking in Dynamic energy savings in cloud-ran: an experimental assessment and cellular radio access networks: potential and challenges. IEEE Commun. implementation (IEEE, Paris, 2017), pp. 791–796 Mag. 53(1), 150–156 (2015) 36. S Samarakoon, M Bennis, W Saad, M Debbah, M Latva-aho, Ultra dense 14. PBerde,MGerola,JHart,YHiguchi,MKobayashi, TKoide,BLantz,B small cell networks: turning density into energy efficiency. IEEE J. Sel. O’Connor, P Radoslavov, W Snow, et al., in Proceedings of the Third Areas Commun. 34(5), 1267–1280 (2016) Workshop on Hot Topics in Software Defined Networking. Onos: towards an 37. N Yu, Z Song, H Du, H Huang, X Jia, Dynamic resource provisioning for open, distributed sdn os (ACM, Chicago, 2014), pp. 1–6 energy efficient cloud radio access networks. IEEE Trans. Cloud Comput. 15. J Medved, R Varga, A Tkacik, K Gray, in World of Wireless, Mobile and (2017). https://ieeexplore.ieee.org/document/7949048/ Multimedia Networks (WoWMoM), 2014 IEEE 15th International Symposium 38. BJ Sahu, S Dash, N Saxena, A Roy, Energy-efficient bbu allocation for on A. Opendaylight: towards a model-driven sdn controller architecture green c-ran. IEEE Commun. Lett. 21(7), 1637–1640 (2017) (IEEE, Sydney, 2014), pp. 1–6 39. D Feng, C Jiang, G Lim, LJ Cimini, G Feng, GY Li, A survey of 16. N McKeown, T Anderson, H Balakrishnan, G Parulkar, L Peterson, J Rexford, energy-efficient wireless communications. IEEE Commun. Surv. Tutor. S Shenker, J Turner, Openflow: enabling innovation in campus networks. 15(1), 167–178 (2013) ACM SIGCOMM Comput. Commun. Rev. 38(2), 69–74 (2008) 40. F Ghods, AO Fapojuwo, FM Ghannouchi, in Electrical and Computer 17. R Mijumbi, J Serrat, J-L Gorricho, N Bouten, F De Turck, R Boutaba, Engineering (CCECE), 2017 IEEE 30th Canadian Conference On.Energy Network function virtualization: state-of-the-art and research challenges. efficiency analysis of a c-ran with distance-based power control (IEEE, IEEE Commun. Surv. Tutor. 18(1), 236–262 (2016) Windsor, 2017), pp. 1–5 18. M Ersue, in Presentation at the IETF# 88 Meeting. Etsi nfv management and 41. T Sigwele, AS Alam, P Pillai, YF Hu, Energy-efficient cloud radio access orchestration—an overview, (Vancouver, 2013) networks by cloud based workload consolidation for 5g. J. Netw. Comput. 19. A de la Oliva, JA Hernandez, D Larrabeiti, A Azcorra, An overview of the Appl. 78, 1–8 (2017) CPRI specification and its application to C-RAN-based LTE scenarios. IEEE 42. Y Zeng, X Wen, Z Lu, Y Chen, H Shao, in Wireless Communication Systems Commun. Mag. 54(2), 152–159 (2016) (ISWCS), 2016 International Symposium On. Joint remote radio head 20. L ETSI, Evolved universal terrestrial radio access (E-UTRA); User equipment activation and beamforming for energy efficient c-ran (IEEE, Poznan, (UE) radio transmission and reception. ETSI TS. 136(101), 16 (2017) 2016), pp. 550–554 21. Y Wu, WY Zou, Orthogonal frequency division multiplexing: a multi-carrier 43. A Li, Y Sun, X Xu, C Yuan, in Computer Communications Workshops modulation scheme. IEEE Trans. Consum. Electron. 41(3), 392–399 (1995) (INFOCOM WKSHPS), 2016 IEEE Conference On. An energy-effective network 22. A Maeder, M Lalam, A De Domenico, E Pateromichelakis, D Wubben, J Bartelt, deployment scheme for 5g cloud radio access networks (IEEE, San R Fritzsche, P Rost, in 2014 European Conference on Networks and Francisco, 2016), pp. 684–689 Communications (EuCNC). Towards a flexible functional split for 44. W Zhao, S Wang, in Vehicular Technology Conference (VTC Spring), 2016 IEEE cloud-RAN networks (IEEE, Bologna, 2014), pp. 1–5 83rd. Remote radio head selection for power saving in cloud radio access 23. M Sawahashi, Y Kishiyama, A Morimoto, D Nishikawa, M Tanno, networks (IEEE, Nanjing, 2016), pp. 1–5 Coordinated multipoint transmission/reception techniques for Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 14 of 15 45. RQ Hu, Y Qian, An energy efficient and spectrum efficient wireless 69. S-H Park, O Simeone, O Sahin, SS Shitz, Fronthaul compression for cloud heterogeneous network framework for 5g systems. IEEE Commun. Mag. radio access networks: signal processing advances inspired by network 52(5), 94–101 (2014) information theory. IEEE Signal Proc. Mag. 31(6), 69–79 (2014) 46. S Parsaeefard, R Dawadi, M Derakhshani, T Le-Ngoc, M Baghani, Dynamic 70. A Checko, AC Juul, HL Christiansen, MS Berger, in Communication resource allocation for virtualized wireless networks in Workshop (ICCW), 2015 IEEE International Conference On. Synchronization massive-mimo-aided and fronthaul-limited c-ran. IEEE Trans. Veh. challenges in packet-based cloud-ran fronthaul for mobile networks (IEEE, Technol. 66(10), 9512–9520 (2017) London, 2015), pp. 2721–2726 47. A Hajisami, D Pompili, Joint virtual edge-clustering and spectrum 71. R Alvizu, G Maier, N Kukreja, A Pattavina, R Morro, A Capello, C Cavazzoni, allocation scheme for uplink interference mitigation in c-ran. Ad Hoc Comprehensive survey on T-SDN: software-defined networking for Netw. 72, 91–104 (2018) transport networks. IEEE Commun. Surv. Tutor. 19(4), 2232–2283 (2017) 48. Z Zhou, J Peng, X Zhang, K Liu, F Jiang, A game-theoretical approach for 72. AC Baktir, A Ozgovde, C Ersoy, How can edge computing benefit from spectrum efficiency improvement in cloud-ran. Mob. Inf. Syst. 2016,(2016). software-defined networking: a survey, use cases, and future directions. https://www.hindawi.com/journals/misy/2016/3068732/abs/ IEEE Commun. Surv. Tutor. 19(4), 2359–2391 (2017) 49. E Zeydan, E Bastug, M Bennis, MA Kader, IA Karatepe, AS Er, M Debbah, 73. M Wollschlaeger, T Sauter, J Jasperneite, The future of industrial Big data caching for networking: moving from cloud to edge. IEEE communication: automation networks in the era of the internet of things Commun. Mag. 54(9), 36–42 (2016) and industry 4.0. IEEE Ind. Electron. Mag. 11(1), 17–27 (2017) 50. A Imran, A Zoha, A Abu-Dayya, Challenges in 5g: how to empower son 74. MR Palattella, M Dohler, A Grieco, G Rizzo, J Torsner, T Engel, L Ladid, with big data for enabling 5g. IEEE Netw. 28(6), 27–33 (2014) Internet of Things in the 5G era: enablers, architecture, and business 51. P Sarigiannidis, A Sarigiannidis, I Moscholios, P Zwierzykowski, Diana: a models. IEEE J. Sel. Areas Commun. 34(3), 510–527 (2016) machine learning mechanism for adjusting the tdd uplink-downlink 75. H Zhang, Y Dong, J Cheng, MJ Hossain, VCM Leung, Fronthauling for 5g configuration in xg-pon-lte systems. Mob. Inf. Syst. 2017,(2017). https:// lte-u ultra dense cloud small cell networks. IEEE Wirel. Commun. 23(6), www.hindawi.com/journals/misy/2017/8198017/abs/ 48–53 (2016) 52. S Han, I Chih-Lin, G Li, S Wang, Q Sun, Big data enabled mobile network 76. N Cvijetic, in Telecommunications Network Strategy and Planning design for 5g and beyond. IEEE Commun. Mag. 55(9), 150–157 (2017) Symposium (Networks), 2014 16th International. Optical network evolution 53. X Zhang, Z Yi, Z Yan, G Min, W Wang, A Elmokashfi, S Maharjan, Y Zhang, for 5g mobile applications and sdn-based control (IEEE, Funchal, 2014), Social computing for mobile big data. Computer. 49(9), 86–90 (2016) pp. 1–5 54. K Zheng, Z Yang, K Zhang, P Chatzimisios, K Yang, W Xiang, Big 77. NJ Gomes, P Chanclou, P Turnbull, A Magee, V Jungnickel, Fronthaul data-driven optimization for mobile networks toward 5g. IEEE Netw. evolution: from cpri to ethernet. Opt. Fiber. Technol. 26, 50–58 (2015) 30(1), 44–51 (2016) 78. Z Pi, F Khan, An introduction to millimeter-wave mobile broadband 55. N Zhang, P Yang, J Ren, D Chen, L Yu, X Shen, Synergy of big data and 5g systems. IEEE Commun. Mag. 49(6), 101–107 (2011) wireless networks: opportunities, approaches, and challenges. IEEE Wirel. 79. TS Rappaport, S Sun, R Mayzus, H Zhao, Y Azar, K Wang, GN Wong, JK Commun. 25(1), 12–18 (2018) Schulz, M Samimi, F Gutierrez, Millimeter wave mobile communications 56. E Hernandez-Valencia, S Izzo, B Polonsky, How will nfv/sdn transform for 5g cellular: it will work!. IEEE Access. 1, 335–349 (2013) service provider opex? IEEE Netw. 29(3), 60–67 (2015) 80. S Rangan, TS Rappaport, E Erkip, Millimeter-wave cellular wireless 57. A Gudipati, D Perry, LE Li, S Katti, in Proceedings of the Second ACM networks: potentials and challenges. Proc. IEEE. 102(3), 366–385 (2014) SIGCOMM Workshop on Hot Topics in Software Defined Networking - HotSDN 81. W Hong, K-H Baek, Y Lee, Y Kim, S-T Ko, Study and prototyping of ’13. SoftRAN: software defined radio access network (ACM Press, New practically large-scale mmwave antenna systems for 5g cellular devices. York, USA, 2013), p. 25 IEEE Commun. Mag. 52(9), 63–69 (2014) 58. IF Akyildiz, P Wang, S-C Lin, SoftAir: a software defined networking 82. T Bai, RW Heath, Coverage and rate analysis for millimeter-wave cellular architecture for 5G wireless systems. Comput. Netw. 85, 1–18 (2015) networks. IEEE Trans. Wirel. Commun. 14(2), 1100–1114 (2015) 59. K Sundaresan, MY Arslan, S Singh, S Rangarajan, SV Krishnamurthy, 83. Y Niu, Y Li, D Jin, L Su, AV Vasilakos, A survey of millimeter wave FluidNet: a flexible cloud-based radio access network for small cells. communications (mmwave) for 5g: opportunities and challenges. Wirel. IEEE/ACM Trans. Netw. 24(2), 915–928 (2016) Netw. 21(8), 2657–2676 (2015) 60. X Foukas, N Nikaein, MM Kassem, MK Marina, K Kontovasilis, in Proceedings 84. M Xiao, S Mumtaz, Y Huang, L Dai, Y Li, M Matthaiou, GK Karagiannidis, of the 12th International on Conference on Emerging Networking E Björnson, K Yang, I Chih-Lin, et al, Millimeter wave communications for EXperiments and Technologies - CoNEXT ’16. FlexRAN: a flexible and future mobile networks. IEEE J. Sel. Areas Commun. 35(9), 1909–1935 programmable platform for software-defined radio access networks (2017) (ACM Press, New York, New York, USA, 2016), pp. 427–441 61. C-Y Chang, N Nikaein, R Knopp, T Spyropoulos, SS Kumar, in 2017 IEEE 85. TS Rappaport, Y Xing, GR MacCartney, AF Molisch, E Mellios, J Zhang, International Conference on Communications (ICC). FlexCRAN: a flexible Overview of millimeter wave communications for fifth-generation (5g) functional split framework over ethernet fronthaul in Cloud-RAN (IEEE, wireless networks—with a focus on propagation models. IEEE Trans. Paris, 2017), pp. 1–7 Antennas Propag. 65(12), 6213–6230 (2017) 62. K Pentikousis, Y Wang, W Hu, Mobileflow: toward software-defined 86. R Taori, A Sridharan, Point-to-multipoint in-band mmwave backhaul for mobile networks. IEEE Commun. Mag. 51(7), 44–53 (2013) 5g networks. IEEE Commun. Mag. 53(1), 195–201 (2015) 63. Bo, H, Youke, W, Chuan’an, W, Ying, W, in 2016 2nd IEEE International 87. M Rebato, M Mezzavilla, S Rangan, F Boccardi, M Zorzi, in European Conference on Computer and Communications (ICCC). The controller Wireless 2016; 22th European Wireless Conference; Proceedings Of. placement problem for software-defined networks (IEEE, Chengdu, 2016), Understanding noise and interference regimes in 5g millimeter-wave pp. 2435–2439 cellular networks (VDE, Oulu, 2016), pp. 1–5 64. Y-C Liang, K-C Chen, GY Li, P Mahonen, Cognitive radio networking and 88. D Huerfano, I Demirkol, P Legg, in Communications Workshops (ICC communications: an overview. IEEE Trans. Veh. Technol. 60(7), 3386–3407 Workshops), 2017 IEEE International Conference On. Joint optimization of (2011) path selection and link scheduling for millimeter wave transport 65. J Wu, P Fan, A survey on high mobility wireless communications: networks (IEEE, Paris, 2017), pp. 115–120 challenges, opportunities and solutions. IEEE Access. 4, 450–476 (2016) 89. S Papaioannou, G Kalfas, C Vagionas, C Mitsolidou, P Maniotis, A Miliou, 66. H Zhang, N Liu, X Chu, K Long, A-H Aghvami, VC Leung, Network slicing N Pleros, in Interactive Mobile Communication, Technologies and Learning. based 5g and future mobile networks: mobility, resource management, 5g small-cell networks exploiting optical technologies with mmwave and challenges. IEEE Commun. Mag. 55(8), 138–145 (2017) massive mimo and mt-mac protocols (Springer, 2017), pp. 805–813. 67. M Peng, C Wang, V Lau, HV Poor, Fronthaul-constrained cloud radio https://link.springer.com/chapter/10.1007/978-3-319-75175-7_79 access networks: insights and challenges. IEEE Wirel. Commun. 22(2), 90. M Sung, S-H Cho, J Kim, JK Lee, JH Lee, HS Chung, Demonstration of 152–160 (2015) 68. C-l I, Y Yuan, J Huang, S Ma, C Cui, R Duan, Rethink fronthaul for soft RAN. ifof-based mobile fronthaul in 5g prototype with 28-ghz millimeter wave. IEEE Commun. Mag. 53(9), 82–88 (2015) J. Light. Technol. 36(2), 601–609 (2018) Pliatsios et al. EURASIP Journal on Wireless Communications and Networking (2018) 2018:136 Page 15 of 15 91. Peng, M, Li, Y, Jiang, J, Li, J, Wang, C, Heterogeneous cloud radio access networks: a new perspective for enhancing spectral and energy efficiencies. IEEE Wirel. Commun. 21(6), 126–135 (2014) 92. M Peng, S Yan, K Zhang, C Wang, Fog-computing-based radio access networks: issues and challenges. IEEE Netw. 30(4), 46–53 (2016) 93. D Liu, B Chen, C Yang, AF Molisch, Caching at the wireless edge: design aspects, challenges, and future directions. IEEE Commun. Mag. 54(9), 22–28 (2016) 94. H Zhang, Y Qiu, K Long, GK Karagiannidis, X Wang, A Nallanathan, Resource allocation in noma based fog radio access networks. IEEE Wirel. Commun (2018). arXiv preprint arXiv:1803.05641. https://arxiv.org/abs/ 1803.05641 95. Y Saito, Y Kishiyama, A Benjebbour, T Nakamura, A Li, K Higuchi, in Vehicular Technology Conference (VTC Spring), 2013 IEEE 77th. Non-orthogonal multiple access (noma) for cellular future radio access (IEEE, Dresden, 2013), pp. 1–5 96. H Zhang, Y Qiu, X Chu, K Long, VC Leung, Fog radio access networks: mobility management, interference mitigation, and resource optimization. IEEE Wirel. Commun. 24(6), 120–127 (2017) 97. T Asai, in 2015 International Conference on Optical Network Design and Modeling (ONDM). 5G radio access network and its requirements on mobile optical network (IEEE, Pisa, 2015), pp. 7–11 98. T Han, X Ge, L Wang, KS Kwak, Y Han, X Liu, 5G Converged cell-less communications in smart cities. IEEE Commun. Mag. 55(3), 44–50 (2017) 99. A Binu, GS Kumar, Virtualization techniques: a methodical review of xen and kvm. Adv. Comput. Commun. 190, 399–410 (2011). https://link. springer.com/chapter/10.1007/978-3-642-22709-7_40

Journal

EURASIP Journal on Wireless Communications and NetworkingSpringer Journals

Published: May 30, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off