Introduction
The deployment of 5G infrastructure presents a significant economic challenge for Mobile Network Operators (MNOs). High costs associated with hardware, site acquisition, and power consumption often make deployment in rural or low-income regions unfeasible, thereby widening the "digital divide." While infrastructure sharing is a recognized cost-saving measure, traditional research often focuses on average coverage or capacity. This paper shifts the focus to Service Continuity, utilizing percolation theory to determine the critical base station (BS) density required to form a "giant connected component"—a continuous signal blanket essential for seamless handovers in a mobile environment.
Infrastructure Sharing Strategies
The authors evaluate three primary infrastructure sharing strategies to understand their impact on network topology and connectivity:
- No Sharing: Users are restricted to the BSs of their own MNO.
- Passive Sharing (Site Sharing): MNOs share physical sites and masts but maintain independent radio equipment and spectrum.
- Active Sharing (Access Sharing): MNOs share the entire Radio Access Network (RAN), allowing users to connect to any available BS regardless of the provider.
Mathematical Modeling and Percolation Theory
To analyze these strategies, the authors combine Stochastic Geometry with Percolation Theory. Base stations are modeled as independent Poisson Point Processes (PPPs). A key innovation is the introduction of an "average coverage radius" ($r_m$). While standard models like the Gilbert disk model assume perfect circular coverage, real-world Signal-to-Interference-plus-Noise Ratio (SINR) coverage is irregular. The proposed radius accounts for global interference, enabling the use of tractable disk-based models to approximate complex SINR-based graphs.
The study identifies a "phase transition"—a tipping point where isolated islands of coverage merge into a single, massive network. A critical finding is the Interference Ceiling: if the product of the SINR threshold ($\beta$) and the interference cancellation factor ($\gamma$) reaches or exceeds 1 ($\beta\gamma \ge 1$), global connectivity becomes physically impossible, regardless of BS density.
Key Results and Performance Gains
The theoretical framework was validated through extensive Monte Carlo simulations in a 4000m × 4000m residential area. The results highlight the dramatic efficiency of active sharing:
- Density Requirements: To achieve a 0.9 percolation probability, No Sharing requires a BS density of $> 1 \times 10^{-6}$ BS/m², whereas Passive Sharing requires $\approx 5 \times 10^{-7}$ BS/m².
- Active Sharing Advantage: Active Sharing requires only $\approx 2.5 \times 10^{-7}$ BS/m². This means MNOs can provide continuous service with 4x fewer base stations compared to non-cooperative deployment.
- Model Accuracy: The "average coverage radius" proved to be a conservative yet highly accurate approximation, ensuring theoretical predictions remain reliable for actual network performance.
Conclusion and Future Directions
This research demonstrates that active sharing significantly lowers the barrier for large-scale service continuity, offering a technical roadmap for bridging the digital divide. While active sharing presents implementation challenges—such as complex roaming protocols and security risks—its topological benefits are undeniable. Future work may extend this model to 6G networks, exploring "dynamic percolation" where network topology adapts in real-time to fluctuating traffic loads and the computational demands of multi-modal AI.
Summary updated