PaperlayerView exact revision on arXiv

Source-linked AI summary

Wireless Backhaul in 5G and Beyond: Issues, Challenges and Opportunities

Berke Tezergil, Ertan Onur

arXiv:2103.08234v2cs.NI

TL;DR

Wireless backhaul is increasingly relevant as 5G technologies and network densification make it more viable, but its efficient use still faces technical challenges. This survey synthesizes developments, challenges, opportunities, applications, security, standardization, and beyond-5G directions. It identifies wireless backhaul as an enabler for applications including rural connectivity, satellite communication, and mobile edge computing.

  • Problem

    Wireless backhaul faces unresolved challenges in resource allocation, deployment, scheduling, power management, coverage, and small-cell deployment, including distributed scheduling for ultra-dense networks.

    Source: §V · p. 10, B. Deployment · p. 15, 3) Open Issues and Challenges · p. 21

  • Method

    The paper surveys wireless-backhaul developments, challenges, enabling applications, security, standardization, and concepts relevant to beyond-5G networks.

    Source: §IX · pp. 44–45

  • Results

    The survey identifies wireless backhaul as an enabler for rural connectivity, satellite backhaul, and mobile edge computing, while highlighting challenges and security issues associated with small-cell multihop backhaul.

    Source: §IX · pp. 44–45

  • Takeaways & Limitations

    The authors conclude that emerging technologies are expected to strengthen wireless backhaul’s role in beyond-5G networks.

    Source: §IX · p. 45

Abstract

from arXiv · show

With the introduction of new technologies such as Unmanned Aerial Vehicle (UAV), High Altitude Platform Station (HAPS), Millimeter Wave (mmWave) frequencies, Massive Multiple-Input Multiple-Output (mMIMO), and beamforming, wireless backhaul is expected to be an integral part of the 5G networks. While this concept is nothing new, it was shortcoming in terms of performance compared to the fiber backhauling. However, with these new technologies, fiber is no longer the foremost technology for backhauling. With the projected densification of networks, wireless backhaul has become mandatory to use. There are still challenges to be tackled if wireless backhaul is to be used efficiently. Resource allocation, deployment, scheduling, power management and energy efficiency are some of these problems. Wireless backhaul also acts as an enabler for new technologies and improves some of the existing ones significantly. To name a few, rural connectivity, satellite communication, and mobile edge computing are some concepts for which wireless backhauling acts as an enabler. Small cell usage with wireless backhaul presents different security challenges. Governing bodies of cellular networks have standardization efforts going on especially for the Integrated Acces-Backhaul (IAB) concept, and this is briefly mentioned. Finally, wireless backhaul is also projected to be an important part of the beyond 5G networks, and newly developed concepts such as cell-free networking, ultra-massive MIMO, and extremely dense network show this trend as well. In this survey, we present the aforementioned issues, challenges, opportunities, and applications of wireless backhaul in 5G, while briefly mentioning concepts related to wireless backhaul beyond 5G alongside with security and standardization issues.

I. INTRODUCTION

Wireless backhaul has become important to 5G as new radio technologies and network densification make it more viable and necessary. This survey reviews its development, challenges, opportunities, applications, security, standardization, and beyond-5G directions.

I. INTRODUCTION

  • Earlier relays and self-backhauled cells saw limited adoption because interference and scheduling remained unresolved, sub-6 GHz spectrum was scarce, densification lagged, and performance trailed fiber.The introduction presents new radio technologies and projected densification as changing these conditions.Source: §I · p. 1

A. Scope

B. Motivation

  • UAVs, HAPS, and mobile or nomadic base stations make wireless backhaul important to future networks, while IAB is envisioned as a major change to network operation.The survey authors also identify a lack of comparative work covering the topic’s perspectives and note that existing surveys are mostly outdated.Source: B. Motivation · p. 2

C. Difference from Other Surveys and Our Contributions

  • The survey distinguishes itself by combining perspectives that prior surveys treat separately, including wireless-backhaul technologies, challenges, applications, security, standardization, and beyond-5G developments.It describes earlier surveys as focused on particular topics, such as small-cell backhaul, IAB, UAVs, HAPS, or satellites.Source: C. Difference from Other Surveys and… · p. 2

D. Outline of the Survey

  • The paper proceeds from backhaul concepts and 5G enablers through research topics and challenges, then covers applications, standardization, security, beyond-5G, and conclusions.The outline assigns security to Section VIII and standardization to Section XIII.Source: D. Outline of the Survey · pp. 2–3

A. An Overview of pre-5G Cellular Networks

  • Cellular backhaul evolved from mixed wired and wireless arrangements in 2G and 3G to LTE’s all-IP architecture, which simplified network design and made backhaul media more interchangeable.The transition from circuit to packet switching and later Carrier Ethernet supported unified voice and data transport.Source: A. An Overview of pre-5G Cellular… · p. 3

B. Wireless Backhaul in pre-5G Networks

Pre-5G relays and self-backhauling offered wireless alternatives but faced spectrum, performance, interference, and scheduling constraints. In 5G, densification increases the need for wireless backhaul, while IAB provides flexible sharing of access and backhaul resources.

B. Wireless Backhaul in pre-5G Networks

  • LTE relays extended coverage but saw limited deployment because scarce sub-6 GHz spectrum made backhaul compete with access, and network densification did not justify the trade-off.Type 1 and Type 2 relays depended on donor eNBs, while their stated scenarios included coverage extension and indoor coverage.Source: B. Wireless Backhaul in pre-5G Networks · pp. 3–4
  • LTE-Advanced self-backhauling let eNBs reuse radio resources for access and backhaul, improving spectrum and cost efficiency but introducing interference and scheduling challenges.An experiment reported coverage and downlink-bitrate improvements despite reduced effectiveness from self-interference and in-band communication.Source: B. Wireless Backhaul in pre-5G Networks · p. 4

A. Network Densification

  • Small-cell densification can reduce deployment costs through wireless backhaul and hardware reuse, but access-backhaul interference and sub-6 GHz performance gaps remain concerns.The passage says mmWave in Release 15 gives wireless backhaul the potential to reach fiber performance, with beam-based links less constrained by access-backhaul interference.Source: A. Network Densification · p. 5

B. Access & Backhaul Resource Management

  • Densification and shared wireless resources create an allocation problem: base stations must divide air-interface resources between user access and backhaul traffic.The network density is described as increasing from 4–5 BS/km2 in 3G to an anticipated 40–50 BS/km2 in 5G, with small cells introduced to realize densification.Source: B. Access & Backhaul Resource Management · p. 5, A. Network Densification · p. 5, §III · p. 5
  • Integrated access and backhaul replaces a fixed resource split with flexible sharing, reuses hardware for access and backhaul, and supports coverage extension and other network functions.The survey also notes IAB’s role in FR2 deployments, alternate paths after unit failure, multiplexing, and multihop transmission; new and unresolved problems remain for extensive use.Source: B. Access & Backhaul Resource Management · pp. 5–6

IV. CURRENT STATE OF TECHNOLOGIES RELATED TO WIRELESS BACKHAUL IN 5G

This section surveys frequency bands, licensing, topologies, and radio technologies relevant to 5G wireless backhaul, describing their capabilities and operating constraints.

IV. CURRENT STATE OF TECHNOLOGIES RELATED TO WIRELESS BACKHAUL IN 5G

  • The section frames frequencies, topologies, MIMO, and duplex modes as design considerations that complement research on problems such as resource allocation and user association.It surveys these technologies before reviewing the literature on wireless-backhaul problems.Source: §IV · p. 6

A. Frequencies

  • Congested lower frequencies motivate higher-band use, where throughput comes from more generous spectrum rather than the complex radio techniques used at microwave frequencies.The section also surveys licensing approaches, including exclusive, unlicensed, and light licensing; 60 GHz coordination may mitigate unlicensed-spectrum problems, while light licensing varies by region.Source: A. Frequencies · p. 6

1) Microwave Frequencies (Sub-40 GHz):

  • Sub-6 GHz backhaul can provide 500–750 Mbps and operate NLoS, while 7–40 GHz can exceed 1 Gbps but requires LoS.The 26–28 GHz range is used in small-cell and point-to-point in-band backhaul deployments, while Ka- and Ku-bands are used in satellite communications.Source: 1) Microwave Frequencies (Sub-40 GHz) · p. 7

2) V-band (57-64 and 64-71 GHz):

  • V-band can offer up to 10 Gbps over distances up to 1 km, but oxygen absorption limits range and operation requires LoS.Its beams cannot penetrate foliage or reflect from other surfaces, and the section also notes comparatively low vendor interest.Source: 2) V-band (57-64 and 64-71 GHz) · p. 7

3) E-band (71-76 and 81-86 GHz):

  • E-band supports 25+ Gbps at around 3 km, with scenario-dependent links reaching 10 Gbps up to 5 miles in suburban settings and 10–20 Gbps up to 2 miles in urban settings.Heavy rainfall can limit its range, though less severely than oxygen absorption affects V-band; E-band also requires LoS.Source: 3) E-band (71-76 and 81-86 GHz) · p. 7

4) W-band (92-110 GHz):

This section surveys higher-frequency options and network techniques for wireless backhaul, outlining their capabilities alongside trade-offs in range, hardware, and operation.

4) W-band (92-110 GHz):

  • W-band offers two subbands of 5 GHz and 7.5 GHz and performs slightly better than D-band in attenuation and range, although D-band is considered the better alternative because of its wider subbands.The section gives D-band contiguous deployments of up to 12.5 GHz and potential throughput of 100 Gbit/s with certain configurations, while noting higher atmospheric attenuation than E-band and the need for beamforming.Source: 4) W-band (92-110 GHz) · p. 7, 5) D-band (110-170 GHz) · p. 8

5) D-band (110-170 GHz):

  • D-band's higher frequencies support potential throughput of 100 Gbit/s with certain configurations, but atmospheric attenuation, beamforming needs, and vendor readiness constrain its use.The band also allows contiguous deployments of up to 12.5 GHz.Source: 5) D-band (110-170 GHz) · p. 8

6) Sub-THz frequencies and beyond:

  • Beyond D-band, free-space optics offers high bandwidth and full-duplex operation but requires LoS and is affected by fog and dust storms.Hybrid FSO/RF systems combine FSO's high capacity and low latency with microwave/mmWave RF's low cost, high coverage, and high availability.Source: 6) Sub-THz frequencies and beyond · p. 8

C. MIMO

  • Massive MIMO uses large antenna arrays to form directional beams and spatially multiplex users, supporting mmWave use, frequency reuse, and service to multiple small cells.The section also describes theoretical energy savings and potential range extension, including reduced power requirements under imperfect CSI.Source: C. MIMO · pp. 8–9

B. Topology

  • PtP uses directional, typically LoS links with dedicated transceivers and full bandwidth, whereas PtMP shares bandwidth across small cells using fewer transceivers and lower setup expenditure.The figure contrasts the link topologies; PtP can suffer where LoS is obstructed, while PtMP can support LoS and NLoS connections.Source: Figure 3 · p. 8, B. Topology · p. 8

D. Half- and Full-duplexing

  • Half-duplex stations transmit or receive at a given time, while beamforming can support full-duplex transmission and reception in different sectors; in-band full duplex remains harder because of interference.Full-duplexing is identified as an open research area for wireless backhauling.Source: D. Half- and Full-duplexing · pp. 9–10

V. CHALLENGES OF WIRELESS BACKHAUL

This section surveys challenges to effective wireless backhaul, emphasizing resource allocation and performance evaluation, as well as optimization methods and architecture-specific concerns in wireless backhaul and related heterogeneous networks.

V. CHALLENGES OF WIRELESS BACKHAUL

  • The survey treats deployment, scheduling, coverage, security, and performance evaluation as distinct challenges, with small-cell traffic scheduling and deployment posing wireless-backhaul-specific problems.Performance evaluations assess field feasibility without performance loss and identify adaptations needed for effective use.Source: §V · p. 10
  • Resource allocation is paramount because scarce bandwidth, energy, links, link capacities, and time slots must be managed dynamically to use wireless backhaul effectively.Allocations of bandwidth or energy must be adjusted to optimize network performance.Source: §V · p. 10, A. Resource Allocation · p. 10

1) State of the Art:

  • [60]’s energy-efficiency methods outperform fixed power allocation, while selective small-cell activation reduces power consumption and low power budgets can improve efficiency.The study models adaptive decoding power and proposes a low-complexity approximation alongside a branch-and-bound global search.Source: 1) State of the Art · p. 11
  • [61] finds load-proportional bandwidth partitioning provides higher coverage than equal partitioning, with an optimal access–backhaul split for each strategy.The IAB-enabled network outperforms the macro-only network only up to a point, after which their performance converges.Source: 1) State of the Art · p. 11
  • [63] jointly allocates backhaul bandwidth and computation for MEC offloading; more UEs reduce offloading, while the algorithm lowers computational overhead and UE transmit power.[64] separately reports that partial offloading outperforms full offloading and local computing in energy efficiency.Source: 1) State of the Art · p. 11
  • [67]’s hierarchical resource-management method for flexible backhaul outperforms five baselines in network utility and converges faster, with similar signaling overhead and little CPU overhead.It separates long- and short-term controls, including routing, flow control, discontinuous transmission, and link scheduling.Source: 1) State of the Art · p. 12
  • [77] uses ARIMA and MLP models trained on synthetic time series based on real network data; all models stay below the MAPE error threshold of 10 and improve capacity planning.The methods also reduce resource wastage.Source: 1) State of the Art · p. 14
  • [78]’s distributed GALA algorithm favors green energy and significantly improves power savings over maximum-rate and α-distributed association, at the cost of negligible latency increases versus α-distributed association.When on-grid power is necessary, GALA balances its use against traffic delivery latency.Source: 1) State of the Art · p. 14
  • Pu et al.’s Markov-approximation algorithm [82] maximizes sum data rate while meeting user demands and SBS backhaul constraints, outperforming OBFD and access-backhaul split in spectral efficiency.Simulations also report slightly higher achieved than required UE data rates and minimized repeating backhaul traffic.Source: 1) State of the Art · p. 14
  • Dai et al. jointly address user association and power allocation in H-CRANs, with lower computational complexity and signaling overhead than centralized max sum-rate optimization.They solve user association centrally and distribute power allocation using variational inequality theory.Source: 1) State of the Art · p. 14
  • Xu et al.’s HetNet survey identifies resource-allocation challenges spanning power, association, bandwidth, interference, fairness, relaying, beamforming, and management of baseband units.The challenges vary by architecture, and integrating IRS and ultra-dense networks also requires careful resource-allocation design.Source: 1) State of the Art · pp. 14–15

2) Lessons Learned:

Effective wireless backhaul in ultra-dense networks depends on adaptable resource allocation, suitable user association, and interference-aware power allocation. Computation-offloading decisions must also account for possible air-interface penalties.

2) Lessons Learned:

  • Dynamic access/backhaul resource splits help networks adapt resource use to changing scenarios, especially in ultra-dense deployments.The section highlights the value of fast algorithms that respond to changing conditions.Source: 2) Lessons Learned · p. 15
  • User association must balance the availability of multiple base stations with small cells’ limited range and line-of-sight requirements.Suboptimal associations may be used to maintain average performance or meet a minimum level.Source: 2) Lessons Learned · p. 15
  • Spatial multiplexing and transmit-power allocation can address UE-side interference, while base-station beamforming combined with power allocation reduces interference effects.Source: 2) Lessons Learned · p. 15
  • Computation-offloading decisions should account for air-interface effects because offloading can cause performance penalties and network-side power wastage.Wireless backhaul can allow offloading computation-heavy tasks, but that does not guarantee a benefit in every case.Source: 2) Lessons Learned · p. 15

3) Open Issues and Challenges:

The section identifies unresolved resource-allocation and deployment challenges, then surveys approaches to placement, routing, and backhaul design. It highlights that network density, mobility, and technology choices shape both performance and implementation complexity.

3) Open Issues and Challenges:

  • 3D networking, spectrum and interference management, and the tradeoff between dynamic allocation and spectrum sensing remain open resource-allocation challenges.New resource-allocation models are also needed to capture the complexity introduced by multiple technologies.Source: 3) Open Issues and Challenges · p. 15
  • Most surveyed works omit mmWave, while jointly considering mmWave, beamforming, and mMIMO makes resource allocation harder and requires further study.Machine-learning allocation also faces limited training resources at a single base station, motivating distributed algorithms or a new training paradigm.Source: 3) Open Issues and Challenges · p. 15

B. Deployment

  • Wireless backhaul makes placement of macro and small cells more flexible, but placement still directly affects network performance.Mobile base stations make deployment dynamic, so finding good placements quickly matters for adapting to changing conditions.Source: B. Deployment · pp. 15–16

1) State of the Art:

  • UAV deployment can raise average sum-rate in congested areas, but may reduce performance for UEs near gNBs and serves many UEs less efficiently.Distributing users among UAVs is key to a high-performance network.Source: 1) State of the Art · p. 17
  • For high-speed trains, faster speeds require UAVs to increase altitude to keep connections up; joint optimization of UAV deployment, caching placement, and user association reaches within 0.02 MOS of exhaustive search,.Source: 1) State of the Art · p. 17
  • In simulated IAB deployments, adding IAB nodes improves average delay, 5th-percentile throughput, and average buffering over donor-only deployment, though all-wired deployment performs best.The cited study also finds wired-first routing better than highest-quality-first for throughput and latency.Source: 1) State of the Art · p. 17
  • UAV deployment studies report gains from learned placement, including reduced blocking and improved capacity, throughput, and resource utilization.One hybrid ARIMA-XGBoost model reports at least 87% prediction accuracy in the worst case; another method reaches 96% of the optimal value without retraining after scenario changes.Source: 1) State of the Art · pp. 17–18
  • Wireless backhaul design studies expose tradeoffs among density, topology, and cost: added density can shift networks from backhaul-limited to interference-limited, while hybrid RF/FSO planning can cost less than fiber-only designs while maintaining reliability and KPIs.Multipath multi-hop backhaul outperforms direct, multihop, and multiple-association alternatives across inter-site distances.Source: 1) State of the Art · pp. 17–19

2) Lessons Learned:

The section distills deployment lessons and open problems for mobile networks, then reviews scheduling research for shared access and backhaul resources. It emphasizes adaptability, scalable learning-based approaches, and tradeoffs among throughput, complexity, and energy use.

2) Lessons Learned:

  • Tree topologies support hierarchical backhaul flow, while mesh topologies offer flexibility and redundancy through multiple links.UAV placement can also give small cells greater freedom and adaptability when deployment is chosen appropriately.Source: 2) Lessons Learned · p. 19
  • Learning-based UAV deployment is presented as a fast, flexible route toward self-organizing networks, compared with optimization approaches such as linear programming.The section notes that current methods are still too slow for large-scale self-organization, and larger-scale learning approaches need improvement.Source: 2) Lessons Learned · p. 19, 3) Open Issues and Challenges · p. 19

3) Open Issues and Challenges:

  • Current UAV research rarely uses mmWave; mmWave deployment remains open, including vertical modeling, beam management, and antenna design.Source: 3) Open Issues and Challenges · p. 19
  • Mobile cells alter mesh and tree topologies, making adaptation essential; the survey views distributed approaches as having an edge in networks with many base stations.Source: 3) Open Issues and Challenges · p. 19

C. Scheduling

  • Scheduling access and backhaul traffic sharing the same medium is essential to approach the network’s performance potential.The survey identifies full-duplex and mmWave as common scheduling research themes and machine learning as a future research direction.Source: C. Scheduling · p. 19

1) State of the Art:

  • A recurrent-network scheduler outperforms max-min, backpressure, DRQN, and r-REINFORCE baselines in packet delay across one- to three-hop users at medium and high loads.Source: 1) State of the Art · p. 19
  • Scheduling studies report tradeoffs: concurrent transmission can outperform MIS-based scheduling at higher energy use, and interference-aware scheduling improves backhaul throughput by 25%.A combined load-and-link-quality policy performs similarly to exhaustive search while being easier to implement; full-duplex gains can eventually encounter weaker backhaul links as a bottleneck.Source: 1) State of the Art · pp. 20–21
  • An edge-coloring scheduler achieves 70%–90% of optimal throughput with 100x better scalability; multiple-RF-chain results reach 85%–90% of optimal using 2%–20% of the time.For full-duplex scheduling, approximation algorithms reach 90% of optimal performance, compared with 60%–70% for half-duplex approximations.Source: 1) State of the Art · p. 21

2) Lessons Learned:

The survey reviews scheduling lessons and open problems, then compares performance-evaluation studies of wireless backhaul across IAB, frequency bands, modulation, deployment, and other settings.

2) Lessons Learned:

  • Spatial multiplexing can double data rate and support full-duplex communication, making effective scheduling central to mmWave performance.Distributed machine learning is promising for ultra-dense scheduling because centralized management incurs too many control messages; interference management without spatial multiplexing remains open.Source: 2) Lessons Learned · p. 21

3) Open Issues and Challenges:

  • Distributed machine learning and interference management without spatial multiplexing remain open scheduling directions for dense and adaptable 5G networks.The survey identifies distributed learning as promising because centralized control messages become excessive with many small cells.Source: 3) Open Issues and Challenges · p. 21

D. Performance Evaluation

  • The performance-evaluation section focuses on studies evaluating concepts themselves, distinct from papers that formulate and solve a problem.These evaluations are included to show the feasibility and gains of wireless backhauling.Source: D. Performance Evaluation · pp. 21–22

1) State of the Art:

  • IAB studies report cell-edge gains alongside possible degradation for users near gNBs, while simulations find IAB can match fiber performance with a small increase in IAB nodes.The comparison with fiber fixes the access/backhaul resource split in simulation; authors anticipate better real-world results with dynamic splitting and planning. An E-band OFDM design also achieves 25 Gbps.Source: 1) State of the Art · pp. 22, 24
  • Frequency-band and gateway studies expose trade-offs: D-band can lead in multi-hop efficiency and throughput, while excessive gateway density can worsen spatial throughput through interference.At 30 GHz, more than 80% of resources go to backhaul versus 45% at 3.5 GHz, whose summed logarithmic user rates are higher; state-of-the-art E-band performs best in the single-link comparison.Source: 1) State of the Art · pp. 22–23
  • For mmWave MIMO-IAB, throughput gains depend on time allocation, user count, RF chains, and antenna assumptions; full-duplex performance is also limited by residual self-interference.The ideal antenna-array case reports a 40% throughput gain and a time-allocation range more than four times wider; interference cancellation and scheduling affect reliability.Source: 1) State of the Art · p. 23
  • Other evaluations cover XGS-PON, self-backhauled small cells, and MANET backhaul, finding technology- and deployment-specific constraints and gains.XGS-PON meets the cited recommendations except upstream nominal line rate and sensitivity; directed antennas improve performance in self-backhauled small cells, and MANET packet loss rises with Wi-Fi interference and hop count.Source: 1) State of the Art · pp. 22–23
  • Single-carrier QAM outperforms OFDM in several evaluated mmWave conditions, while a hybrid RF/FSO link uses RF when visibility falls below 50 m.SC-QAM performs better for higher-order modulation BER, 64QAM phase noise, and faulty antenna elements; dust affects FSO more than RF.Source: 1) State of the Art · p. 23

2) Lessons Learned:

The survey presents IAB as promising, notes mmWave limitations and the potential of hybrid media, and identifies evaluation gaps for higher bands and beyond-5G concepts.

2) Lessons Learned:

  • IAB is presented as promising for future use, with performance described as adequate for daily use and compatible with an integral role in 5G.Source: 2) Lessons Learned · p. 24
  • mmWave offers throughput but has shorter range than sub-6 GHz and greater sensitivity to weather, path loss, and blockage.The survey names deployment strategies, beamforming, intelligent reflecting surfaces, and mMIMO as ways that may boost performance.Source: 2) Lessons Learned · p. 24
  • Combining communication technologies may yield more robust systems than individual media, as shown for hybrid FSO/mmWave approaches.Source: 2) Lessons Learned · p. 24

3) Open Issues and Challenges:

  • Performance evaluation of IAB at higher bands such as V-, E-, and D-band remains an open research area.Source: 3) Open Issues and Challenges · p. 24
  • Cell-free networking, integrated terrestrial-satellite networks, and THz communications lack the performance evaluations described for older concepts.The survey argues that evaluation could hasten research and draw interest to these beyond-5G areas.Source: 3) Open Issues and Challenges · p. 24

E. Coverage

  • In 5G small-cell deployments, coverage performance is directly linked to wireless-backhaul quality.Source: E. Coverage · p. 24

1) State of the Art:

The survey reviews coverage methods for relayed, small-cell, and aerial networks, then highlights tradeoffs and open challenges in extending coverage. It emphasizes deployment, rural coverage, and updated modeling as continuing concerns.

1) State of the Art:

  • Flexible TDD outperforms Split TDD despite higher interference, while relay nodes improve 5th-percentile downlink and uplink throughput.The surveyed work also compares IBFD and FDD: FDD has better coverage probability, whereas IBFD has higher rates but suffers from inter-tier interference and low coverage.Source: 1) State of the Art · p. 25
  • D2D relays can expand coverage, but spectral efficiency eventually falls as backhaul-link deterioration outweighs the benefit of serving more UEs.Full-duplex amplify-and-forward relays improve coverage probability and transmission capacity, though transmission capacity declines beyond a certain relay density as interference grows.Source: 1) State of the Art · p. 25
  • UAV-based coverage outsourcing combines reputation-based auctions with blockchain and SVM monitoring; moving UAVs toward underserved areas improves service quality while worsening service elsewhere.More monitoring nodes increase true-classification probability and reduce false alarms; higher UAV cost for deviation is positively correlated with ergodic capacity.Source: 1) State of the Art · p. 25

2) Lessons Learned:

  • The survey identifies small cells as the leading current approach to cell-edge coverage, with mmWave, beamforming, and mMIMO supporting dense deployments while mitigating interference.The survey also describes machine learning as promising for coverage-hole and interference mitigation, but says faster responses to changing conditions require further research.Source: 2) Lessons Learned · p. 26

3) Open Issues and Challenges:

  • Coverage gains can come at the expense of capacity or throughput, leaving coverage-capacity balance an open problem for future cellular networks.The tradeoff is linked to the better range but lower capacity of lower frequencies.Source: 3) Open Issues and Challenges · pp. 26–27
  • Network deployment design, especially small-cell placement, is closely tied to coverage gains.Source: 3) Open Issues and Challenges · p. 27
  • Ubiquitous coverage that meets 5G demands remains difficult, particularly in rural settings where coverage and performance challenges persist.Source: 3) Open Issues and Challenges · p. 27
  • Coverage models need to move beyond 2D for mmWave and aerial or satellite components, while infrastructure sharing remains an avenue for improving coverage quality and availability.The survey calls for 3D models and further study of network slicing and other technologies for effective infrastructure sharing.Source: 3) Open Issues and Challenges · p. 27

F. Security Aspects of Wireless Backhaul

Wireless backhaul security must address an exposed radio medium while balancing protection against cost, latency, and network performance. The survey reviews security requirements and related work, then identifies lightweight and distributed security as open research directions.

F. Security Aspects of Wireless Backhaul

  • Security measures can increase CapEx and OpEx and reduce network performance, so operators face a tradeoff between protection and service performance.Source: F. Security Aspects of Wireless Backhaul · p. 27
  • Wireless backhaul exposes transmitted data to eavesdropping and man-in-the-middle attacks, making confidentiality, integrity, and mutual authentication key requirements.Authentication also regulates network access and helps prevent unwanted users from connecting.Source: F. Security Aspects of Wireless Backhaul · p. 27
  • Session-key cryptography must protect transmissions and remain fast enough to limit added latency, especially for eMBB and URLLC.Fresh keys are also needed to help prevent replay and spoofing attacks, and handovers or migrations must be secured.Source: F. Security Aspects of Wireless Backhaul · pp. 27–28

1) State of the Art:

  • The reviewed security work calls for mutual authentication, perfect forward secrecy, and lightweight cryptography; it also identifies PUFs as a way to strengthen cryptographic randomization.The survey notes that operators must account for disruption to their businesses when selecting security implementations.Source: 1) State of the Art · p. 28
  • Related work describes a handover-security protocol that satisfies confidentiality, integrity, authentication, key management, and perfect forward secrecy while reducing computational overhead relative to existing 5G security protocols.Source: 1) State of the Art · p. 28

2) Open Issues and Challenges:

The survey identifies security in dense and distributed networks as an open challenge, particularly where performance and protection must be balanced. It also reviews wireless-backhaul applications and aerial-network research, while noting scope boundaries and operational constraints.

2) Open Issues and Challenges:

  • As network density and attack surface grow, security must not be traded away for lower latency, greater capacity, or better connection quality.Lightweight cryptography is identified as an open research area for meeting performance requirements while increasing security.Source: 2) Open Issues and Challenges · p. 29
  • Distributed security and SDN/NFV policies may reduce security-decision latency, but distributed management also creates a larger attack surface.Secure caching is needed alongside performance-oriented caching approaches.Source: 2) Open Issues and Challenges · p. 29

VI. OPPORTUNITIES AND APPLICATIONS

  • The applications section links wireless backhaul with mobile edge computing, rural connectivity, and satellite backhaul, which it describes as becoming more competitive with 5G advances.It says wireless backhaul provides an opportunity to bring edge nodes closer to base stations.Source: §VI · p. 29

A. HAPS and UAV Usage in the Network

  • The HAPS/UAV section is not a comprehensive survey of all related works; it focuses on the technology itself and omits works more closely tied to other topics.Source: A. HAPS and UAV Usage in the Network · p. 30
  • HAPS and UAVs can change location, maintain line of sight at high altitude, and move closer to UEs, but their use raises placement, channel-modeling, and resource-management challenges.These platforms may act as mobile network components and enable lower-power transmission by moving closer to users.Source: A. HAPS and UAV Usage in the Network · p. 30

1) State of the Art:

  • One reviewed optimization reports a 47% network-profit increase when altitude, power, subcarrier allocation, and user association are jointly optimized.The reported gains are more evident in suburban and urban settings and decrease in dense, high-rise urban areas; two MABS and two SABS almost double profit versus one ground base station.Source: 1) State of the Art · p. 30
  • In caching-enabled UAV networks, added UAV density improves successful content delivery only up to a point, after which line-of-sight interference reduces it.The survey also reviews UAV-enabled backhaul, HAPS use cases, and learning-based non-terrestrial access and scheduling approaches.Source: 1) State of the Art · pp. 30–32
  • For multi-ABS trajectory optimization, backhaul capacity is a bottleneck until it rises enough for maximum transmit power to become limiting.Increasing maximum flight time raises max-min data rate, while constraining propulsion power increases service time.Source: 1) State of the Art · pp. 31–32

2) Open Issues and Challenges:

The section surveys open integration challenges and research on satellite backhauling, from aerial-network design and satellite-terrestrial coordination to rural and edge applications. Satellite backhaul offers broad coverage and rapid deployment, but cost, regulation, latency, and system maturity constrain its use.

2) Open Issues and Challenges:

  • UAV/HAPS integration still faces regulatory and frequency constraints, diverse use-case requirements, resource-management demands, stability and mobility issues, and gaps in channel models and evaluation.The section also identifies low-overhead management and seamless switching across URLLC and mMTC as challenges.Source: 2) Open Issues and Challenges · p. 32

B. Satellite Backhauling

  • Satellite backhaul offers rapid deployment and extensive coverage, but high ownership and transport costs, differing national rules, and unresolved stakeholder responsibilities constrain its use.The passage gives satellite data-transport costs of up to $350 per Mbit as of 2019.Source: B. Satellite Backhauling · p. 33
  • Satellite research spans LEO constellations and integrated space-air-ground networks, where satellite coverage, aerial capacity, and dense ground access must be coordinated.The section describes LEO constellations and notes that combining segment strengths also combines their challenges.Source: B. Satellite Backhauling · p. 33
  • New wireless technologies have renewed interest in satellite backhaul, but its role in future networks depends on technologies maturing for commercial use.The section also identifies non-terrestrial networks as groundwork for beyond-5G networks.Source: B. Satellite Backhauling · p. 33, 1) State of the Art · p. 37

1) State of the Art:

  • Satellite backhaul research also covers LoRa connectivity, UAV-satellite resource sharing, IoT offloading, and satellite-supported edge computing and live streaming.The surveyed examples include indirect satellite access for LoRa and context-aware content delivery using multicast when needed.Source: 1) State of the Art · pp. 35–36
  • Proposed satellite-terrestrial solutions use joint carrier assignment, resource management, beamforming, and hybrid network reconfiguration to manage interference and improve network performance.The surveyed work also considers trade-offs between computational effort and performance in beamforming designs.Source: 1) State of the Art · pp. 34–35
  • Experimental and architectural work demonstrates satellite integration with 5G networks, including a mobile SATCOM-equipped van that can establish a standalone cell when satellite line of sight is available.After shadowing in the van demonstration, reconnection took a few seconds to a minute.Source: 1) State of the Art · p. 35

2) Open Issues and Challenges:

The section describes satellite and rural connectivity as opportunities with unresolved technical, economic, and operational constraints. Across these settings, system integration, resource management, coverage economics, and sustainable business models remain central concerns.

2) Open Issues and Challenges:

  • SAGIN deployment must coordinate hierarchy, network elements, communication technologies, and aerial or space-element trajectories, while satellite and UAV/HAPS routing and mobility coordination remain open problems.The section notes that existing routing protocols cannot be used for vertical systems and that satellite and UAV/HAPS mobility differences can cause disconnections.Source: 2) Open Issues and Challenges · p. 37
  • Satellite-network challenges include managing LEO/GEO delay, enabling high-performance satellite base stations, and coordinating spectrum and interference across orbits.Direct satellite service requires onboard L1, L2, and L3 processing and power management for low-latency operation.Source: 2) Open Issues and Challenges · p. 37
  • Satellite resource management must rapidly meet QoS requirements, provide service-area coverage, and manage mobile links; automation is desirable but not mandatory.The passage frames development of such systems as an open challenge for satellite networks.Source: 2) Open Issues and Challenges · p. 37

C. Rural Connectivity

  • Rural coverage remains uneven: in 2017, 3.8 billion people were offline and 1.2 billion lacked broadband, with most unconnected people living in rural areas of developing countries.The passage also reports 2019 internet penetration of 89.4% in North America and 37.3% in Africa.Source: C. Rural Connectivity · p. 37
  • Rural deployments face sparse populations, weak returns, costly infrastructure, and limited power or fiber access; low population density also makes operator-government cooperation and new business models important.Deployment infrastructure can cost twice as much while revenue opportunities can be up to ten times lower than in urban settings.Source: C. Rural Connectivity · pp. 37–38
  • Rural backhaul has historically relied on microwave or satellite, while newer mmWave and FSO options mainly bring backhaul to rural areas rather than directly solving coverage and range from the core.The section identifies high satellite latency and initial cost, and microwave repeater and spectrum costs, as shortcomings of earlier approaches.Source: C. Rural Connectivity · pp. 37–38
  • Rural energy and backhaul choices affect power use: small-cell networks with microwave or satellite backhaul raise consumption, whereas LR-PON has negligible impact in the surveyed analysis.The comparison considers rural access and backhaul configurations over 2010–2021.Source: C. Rural Connectivity · p. 38
  • Rural proposals include UAV-based 5G coverage, which was reported as 2.5 times cheaper than a traditional 5G solution, and RuralPlan, which used 32 kWh versus MaxCov's 86 kWh in a limited-time solution.The section concludes that technical advances alone do not ensure commercial feasibility; new CapEx/OpEx models or other concepts are needed.Source: C. Rural Connectivity · p. 39

D. Mobile Edge Computing

The survey reviews MEC’s relationship to wireless backhaul, including offloading, caching, and resource-allocation work, then outlines IAB standardization architectures and requirements. Across these topics, it highlights deployment and coordination choices alongside performance trade-offs.

D. Mobile Edge Computing

  • MEC moves computation-heavy tasks toward the network edge, but deployment location, generated data volume, and offloading decisions remain challenges.Offloading spends air resources to preserve UE computation and power.Source: D. Mobile Edge Computing · p. 39
  • Startup delay fell from 10–20s to 2.6–2.7s in the satellite-backhauled MEC streaming framework.Holding more video segments improves aggregated client throughput while reducing single-link throughput; longer segments increase latency but require fewer held segments.Source: D. Mobile Edge Computing · pp. 39–40
  • Tan et al.’s distributed ADMM approach beats caching-only and FD-only schemes, but slightly trails its centralized version in system utility and cost.The approach jointly considers user association, power control, caching, computing offloading, and resource allocation.Source: D. Mobile Edge Computing · p. 40
  • Pham et al.’s distributed offloading and resource-allocation algorithm outperforms HODA in percentage of offloading users and computation overhead.The framework jointly addresses user association, subchannel assignment, transmit power, and MEC-server computation resources.Source: D. Mobile Edge Computing · p. 40

VII. STANDARDIZATION EFFORTS ON WIRELESS BACKHAUL

  • For 5G NR, IAB is framed as a way to reduce wired-network dependence, while dense-urban small-cell scenarios and PtMP links feature in broader backhaul planning.The cited scenario discussion considers V-band to D-band frequencies, function splits, and centralized or distributed RAN.Source: §VII · pp. 40–41
  • IAB standardization considers CU/DU-split architectures and nested-tunneling connections from IAB nodes to wired IAB donors.The study item also covers spanning-tree and DAG topologies, with DAG link and route redundancy used to increase reliability.Source: §VII · p. 41
  • The IAB deployment recommendation covers in-band and out-of-band operation, multihop backhaul, topology adaptation, and SA and NSA modes.The cited study item also covers link discovery, scheduling, resource allocation, synchronization, and interference management.Source: §VII · p. 41

VIII. WIRELESS BACKHAUL BEYOND 5G

The survey examines beyond-5G backhaul prospects in higher frequency bands, dense networks, IAB, and ultra-massive MIMO. It describes potential gains alongside open technical and performance limitations.

A. Frequency Use Beyond 5G

  • THz offers high bandwidth and low latency, but its range is reduced by free-space loss and effective use awaits mature transceivers and photonics.The survey identifies indoor and fixed uses such as data centers and VR/AR; outdoor or cellular THz would require mMIMO and beamforming.Source: A. Frequency Use Beyond 5G · p. 41

B. Extremely Dense Networks

  • All simulated moving-access-point configurations improve over the no-moving-access-point baseline, with significantly reduced outage probability and less need for stationary APs.The latter can reduce operator expenditure, while moving elements introduce blockage, reliability, availability, and distributed-management challenges.Source: B. Extremely Dense Networks · p. 42
  • Extremely dense networks still face open service-requirement and management challenges, including signaling overhead from many connected elements.The discussion notes high-reliability, low-latency, high-data-rate requirements in automation and manufacturing.Source: B. Extremely Dense Networks · p. 42

C. Integrated Access and Backhaul

  • IAB can provide coverage and installation-flexibility advantages over fiber, but it still has worse performance, even with mmWave.FSO and hybrid FSO/RF are discussed as possible ways to improve IAB performance, with hybrid links offering reliability and atmospheric-effect tolerance.Source: C. Integrated Access and Backhaul · p. 42

D. Ultra-Massive MIMO

  • Ultra-massive MIMO’s larger arrays can boost signal strength by steering and focusing beams, supporting small-cell effectiveness at sub-6 GHz and higher bandwidth at mmWave and above.The survey identifies cost- and power-efficient array design as a challenge, and says channel models, empirical evaluation, and physical implementation remain open research areas.Source: D. Ultra-Massive MIMO · pp. 42–43

E. Integrated Space and Terrestrial Networks

The survey considers beyond-5G integration of terrestrial, airborne, and space networks, cell-free access points, and intelligent reflecting surfaces. It describes their possible backhaul roles while noting deployment and coordination challenges.

E. Integrated Space and Terrestrial Networks

  • Integrating terrestrial and space networks through an airborne layer could support global coverage, but current aerial microwave links lack the bandwidth for the envisioned demands.The survey identifies 3D networking and current mmWave limitations—including communication distance and spatial multiplexing—as challenges.Source: E. Integrated Space and Terrestrial… · p. 43

F. Cell-Free Networking

  • Cell-free networking replaces a single massive antenna array with cooperating low-cost access points and is reported to improve 95%-likely per-user throughput by five to ten times over small cells.Wireless backhaul could help these access points operate more effectively in ultra-dense networks.Source: F. Cell-Free Networking · p. 43
  • Cell-free operation raises beam-selection and initial-access challenges, while coordination and delegation to CPUs trade improved performance against complexity and fronthaul signaling.Machine learning is proposed for beam selection; removing cell structure also complicates AP identification and synchronization.Source: F. Cell-Free Networking · pp. 43–44
  • Cell-free research still has open questions around user limits, AP placement, and effects of random scatterers and locations.Uplink transmission, multi-connectivity, heterogeneous radios, and the volume of AP–CPU data exchanges are also identified as unresolved challenges.Source: F. Cell-Free Networking · p. 44

G. Intelligent Reflecting Surfaces

  • IRS-assisted LoS links are sensitive to geometry: a 3.5° change reduces TX/RX power radiation patterns by 70%.In wireless-backhaul meshes, more IRS reflectors improve signal error rate and reduce outage probability; too few reflectors make increased hop count infeasible.Source: G. Intelligent Reflecting Surfaces · p. 44
  • Proposed IRS deployments include UAV swarms and tethered balloons for reflected paths, and HAPS-mounted surfaces as a low-power alternative to relays.The tethered-balloon design can serve as a low-cost access point; HAPS-mounted surfaces are described as supporting full-duplex communication.Source: G. Intelligent Reflecting Surfaces · p. 44
  • IRS research challenges include channel sensing and estimation without dedicated hardware, practical transmitter-to-surface protocols, and fast reconfiguration in HetNets.Machine-learning passive beamforming and IRS-assisted D2D, mmWave, and THz communications are also identified as research topics.Source: G. Intelligent Reflecting Surfaces · p. 44

IX. CONCLUSION

  • The survey concludes that wireless backhaul’s challenges include resource allocation, scheduling, coverage, power management, and deployment, while beyond-5G technologies are expected to strengthen its role.It also reviews security issues for small-cell multihop backhaul and standardization efforts, particularly IAB.Source: §IX · pp. 44–45
Loading 2103.08234v2…