Probabilistic Caching for Small-Cell Networks With Terrestrial and Aerial Users

Fei Song, Jun Li, Ming Ding, Long Shi, Feng Shu, Meixia Tao, Wen Chen, H. Vincent Poor

TL;DR

This paper tackles the challenge of efficiently caching popular content in ultra-dense small-cell networks that serve both ground-based (terrestrial) and aerial (UAV) users. They introduce a novel "optimized caching strategy" (OCS) that significantly improves the successful download probability (SDP) compared to existing methods, especially in dynamic network conditions.

What Problem Does This Paper Address?

The explosion of mobile data traffic, driven by video streaming and new applications, is putting immense pressure on cellular networks. 5G networks are increasingly incorporating ultra-dense small cells (UD-SCNs) and aerial users (UAVs) to boost capacity. A key strategy to alleviate backhaul congestion and improve user experience is "wireless caching," where popular content is pre-stored in base stations (SBSs). However, efficiently deciding what to cache and where is complex, especially with the added mobility and unique propagation characteristics of UAVs, and the dynamic nature of SBSs (turning on/off as needed). This paper aims to maximize the probability that users can successfully download their requested files in such a dynamic, two-tier (terrestrial and aerial) network environment.

Methods

The authors employ a sophisticated approach combining stochastic geometry and advanced mathematical analysis.

  • Stochastic Geometry: They model the locations of SBSs, terrestrial users (TUs), and aerial users (UAVs) as independent Poisson Point Processes (HPPPs). This allows for rigorous analysis of network performance under random deployments.
  • Sophisticated Path Loss Model: They utilize a path loss model that accounts for both line-of-sight (LoS) and non-line-of-sight (NLoS) transmissions, crucial for accurately modeling signal propagation between ground users, aerial users, and SBSs, and incorporating the impact of height differences.
  • Dynamic On-Off Architecture: The paper considers an SBS architecture where base stations are only active when serving users, which is more energy-efficient and relevant for future 5G deployments.
  • Probabilistic Caching: Instead of deterministic placement, they focus on probabilistic caching, where each SBS caches files with a certain probability. This offers greater flexibility and robustness in dynamic networks.
  • Optimized Caching Strategy (OCS): The core contribution is the development of an OCS that aims to maximize the average successful download probability (SDP) by optimizing the caching probability for each file. This is formulated as an optimization problem and solved using a partial Lagrangian method.
  • Performance Analysis: They derive analytical expressions for the SDP under different caching strategies (OCS, Popular Caching Strategy - PCS, and Uniform Caching Strategy - UCS) and analyze the impact of various network parameters.

Datasets

Not discussed in the paper. The authors rely on analytical modeling and simulations based on parameters derived from 3GPP standards and existing literature.

Key Insights and Ideas

  • OCS Superiority: The proposed Optimized Caching Strategy (OCS) consistently outperforms both the Popular Caching Strategy (PCS) and the Uniform Caching Strategy (UCS) across various scenarios, demonstrating its effectiveness in maximizing user download success.
  • Dynamic Network Benefits: The dynamic on-off architecture for SBSs, combined with probabilistic caching, offers advantages over traditional always-on architectures, particularly in dense networks.
  • Trade-offs in Caching Strategies:
    • PCS vs. UCS: The choice between PCS and UCS depends heavily on network density and file popularity. PCS is better in sparse networks or when a few files are overwhelmingly popular (high Zipf exponent). UCS performs better in dense networks or when file popularity is more evenly distributed.
  • Cache Size Impact: Increasing cache size generally improves SDP for all strategies, but the OCS maintains its lead.
  • UAVs Add Complexity: UAVs introduce unique challenges due to their mobility and distinct propagation characteristics, requiring specialized path loss models. Their height significantly impacts their download success, but not that of terrestrial users.
  • Density is Key: Higher SBS density generally leads to better SDP, but the performance gains eventually saturate. The crossover point where UCS outperforms PCS shifts with cache size and density.

Key Results

  • The OCS consistently achieves higher average SDP than PCS and UCS.
  • The paper provides analytical expressions for SDP under different caching strategies and network conditions.
  • Simulations validate the analytical results, showing good agreement between theoretical predictions and practical performance.
  • The study quantifies the impact of SBS density, cache size, file popularity distribution (Zipf exponent), and UAV height on the overall network performance. For instance, they show how the crossover point between PCS and UCS performance shifts with increasing cache size.

Limitations and Future Work

  • Uplink Performance: The current analysis focuses on downlink performance. Analyzing uplink performance, especially with terrestrial and aerial users acting as interferers, is noted as a challenging but important future direction.
  • Coded Caching: The paper uses uncoded caching. Exploring the benefits of coded caching strategies to further enhance performance is suggested as future work.
  • 3D HPPP for UAVs: The current model assumes UAVs at a fixed height. Optimizing caching probabilities for UAVs deployed as a 3D HPPP is a complex problem left for future research.
  • Different Request Probabilities: While the paper briefly discusses scenarios with different request probabilities for TUs and UAVs, a more in-depth analysis could be beneficial.

Summary updated

Abstract

The support for aerial users has become the focus of recent 3GPP standardizations of 5G, due to their high maneuverability and flexibility for on-demand deployment. In this paper, probabilistic caching is studied for ultra-dense small-cell networks with terrestrial and aerial users, where a dynamic on-off architecture is adopted under a sophisticated path loss model incorporating both line-of-sight and non-line-of-sight transmissions. Generally, this paper focuses on the successful download probability (SDP) of user equipments (UEs) from small-cell base stations (SBSs) that cache the requested files under various caching strategies. To be more specific, the SDP is first analyzed using stochastic geometry theory, by considering the distribution of such two-tier UEs and SBSs as Homogeneous Poisson Point Processes. Second, an optimized caching strategy (OCS) is proposed to maximize the average SDP. Third, the performance limits of the average SDP are developed for the popular caching strategy (PCS) and the uniform caching strategy (UCS). Finally, the impacts of the key parameters, such as the SBS density, the cache size, the exponent of Zipf distribution and the height of aerial user, are investigated on the average SDP. The analytical results indicate that the UCS outperforms the PCS if the SBSs are sufficiently dense, while the PCS is better than the UCS if the exponent of Zipf distribution is large enough. Furthermore, the proposed OCS is superior to both the UCS and PCS.
BibTeX
@article{song2019probabilistic,
  author = {Fei Song and Jun Li and Ming Ding and Long Shi and Feng Shu and Meixia Tao and Wen Chen and H. Vincent Poor},
  title = {{Probabilistic Caching for Small-Cell Networks With Terrestrial and Aerial Users}},
  journal = {IEEE Trans. Vehicular Technology},
  year = {2019},
  doi = {10.1109/tvt.2019.2929839},
}

Loading related content...

Figure