Abstract
Benefiting from the usage of the high-frequency band, utilizing part of the large available bandwidth for wireless backhauling is feasible without considerable performance sacrifice. In this context, integrated access and backhaul (IAB) was proposed by 3GPP to reduce the fiber optics deployment cost of 5G and beyond networks. In this paper, we first give a brief introduction of IAB based on the 3GPP release. After that, we survey existing research on IAB networks, the integrations of IAB to cache-enabled network, optical communication transport network, and the non-terrestrial network. Finally, we discuss the challenges and opportunities that might arise while developing and commercializing IAB networks.
The Challenge of Network Densification
As we move toward 5G-Advanced and 6G, network "densification"—deploying a massive number of small-cell base stations—is the only way to meet skyrocketing data demands. However, the "umbilical cord" of the cellular network is the backhaul: the connection from the base station to the core network.
Traditionally, this requires optical fiber, which is prohibitively expensive, costing between $100 and $200 per meter in urban environments due to trenching and installation. IAB solves this by allowing base stations to "self-backhaul" wirelessly. By utilizing the massive bandwidth available in millimeter-wave (mmWave) bands, a base station can serve local users while simultaneously acting as a relay for other base stations, eliminating the need for a fiber connection at every single site.
Survey Methodology and Roadmap
This survey categorizes and synthesizes hundreds of recent studies to provide a comprehensive roadmap of the IAB landscape. The research is structured into three primary pillars:
- Standardization: An overview of 3GPP Release 16 and 17, focusing on the architecture split between Central Units (CU) and Distributed Units (DU), and the introduction of the Backhaul Adaptation Protocol (BAP) for routing.
- Analytical Modeling: The use of Stochastic Geometry (SG) to characterize coverage probability and rate distributions in large-scale, random deployments.
- Optimization Techniques: A review of how researchers apply Game Theory, Reinforcement Learning (RL), and Convex Optimization to solve complex problems like interference mitigation, power allocation, and multi-hop routing.
Key Insights and Ideas
- The In-Band vs. Out-of-Band Trade-off: In-band IAB (sharing the same frequency for access and backhaul) offers the highest spectral efficiency but introduces "self-interference," where a node's own transmission can drown out the signal it's trying to receive.
- The Multi-hop Bottleneck: Unlike traditional cellular networks, IAB supports multi-hop relaying. However, the end-to-end rate is limited by the "weakest link" in the chain. Routing is no longer just a networking problem but a physical-layer optimization problem.
- Caching as a Backhaul "Relief Valve": By storing popular content directly at the IAB node (edge caching), the network can reduce backhaul traffic by up to 48%, significantly lowering latency for the end-user.
- Non-Terrestrial Extensions: IAB can be integrated into UAVs (drones) and LEO satellites to provide "coverage from the sky" in rural or disaster-stricken areas, extending connectivity beyond traditional terrestrial limits.
Quantitative Results
The survey highlights several quantitative milestones from the literature:
- Throughput Gains: Field trials showed that deploying even a single IAB node can improve system throughput by 81.5% compared to non-IAB setups.
- AI Superiority: Reinforcement Learning-based schedulers achieved 2.6 times the gain in transmission delay reduction compared to traditional max-min or backpressure schedulers.
- UAV Coverage: Integrating UAVs as mobile IAB nodes can improve downlink coverage by 31% by optimizing their 3D hovering positions to avoid blockages.
- Virtualization: Applying wireless network virtualization to IAB can result in a 60% increase in average total utility for service providers.
Limitations and Future Work
Several "open frontiers" remain for future research:
- Mobility Management: Most current IAB models assume static base stations. In reality, if an IAB node is on a moving vehicle, the backhaul topology must be recalculated in milliseconds.
- 3D Realism: Many analytical models simplify the world into 2D planes. Future work needs to incorporate 3D terrain and building height data to accurately model mmWave blockages.
- Intelligent Self-Organization: There is a need for "Intelligent IAB" controllers that can sense real-time traffic demand and autonomously reconfigure the backhaul mesh to prevent congestion.
Summary updated