Optimizing Content Placement and Delivery in Wireless Distributed Cache Systems Through Belief Propagation

Jianbin Chuan, Bo Bai, Xuewei Wu, Hongming ZhangIEEE Access 2020 · OpenReview

Abstract

Wireless distributed cache system (WDCS) is a promising technique to offload traffic and improve energy efficiency in content-centric networks. In this paper, the content sharing problem is considered by minimizing the average energy consumption for content caching and delivering in device-to-device (D2D) enabled cellular networks. The problem is formulated as a joint optimization problem of multiple interactive variables, which is NP-hard. In order to solve this problem, we first decompose the original problem into three subproblems: content helper (CH) selection, content caching, and content delivering. Specifically, in CH selection, we propose a PageRank-based distributed algorithm to estimate the centrality of mobile users (MUs) by jointly analyzing their social relationships and channel conditions. Based on the selected CHs, content caching and content delivering are formulated into two weighted matching problems. Two belief propagation (BP) based distributed algorithms are then proposed by considering content popularity and channel states. Furthermore, a heuristic-based alternating iterative optimization method (HAIOM) is proposed to improve the average energy consumption performance. Simulation results show that the proposed scheme reduces average energy consumption and delivery delay while increasing the caching hit ratio.

Introduction

As mobile data traffic continues to grow exponentially, traditional base stations (BSs) face significant congestion. Caching popular content directly on mobile users' (MUs) devices via Device-to-Device (D2D) communication offers a viable solution to offload traffic and enhance energy efficiency. However, managing a Wireless Distributed Cache System (WDCS) involves complex, interconnected decisions: identifying optimal Content Helpers (CHs), determining what content to cache under storage constraints, and efficiently delivering that content to requesters. This joint optimization is NP-hard due to the high degree of interaction between these variables.

Methodology: The HAIOM-BPBP Framework

To address this complexity, the proposed framework decomposes the problem into three sequential yet interactive subproblems: CH selection, content caching, and content delivery.

Figure 1: System model and the three-stage optimization framework for content placement and delivery in WDCS.

1. Socially-Aware Helper Selection

Traditional selection methods often focus solely on signal strength. This paper introduces a PageRank-based distributed algorithm that evaluates a user's "centrality." By analyzing both physical channel quality and social interaction patterns (e.g., encounter frequency), the algorithm identifies MUs who are most likely to be in proximity to multiple potential requesters.

2. Belief Propagation for Caching and Delivery

Once CHs are selected, the system must optimize what they store and how they serve others:

  • Content Caching: Modeled as a weighted many-to-many matching problem.
  • Content Delivery: Modeled as a one-to-many matching problem.

The authors employ Belief Propagation (BP) algorithms for these tasks. BP allows nodes to coordinate locally by passing messages regarding their preferred states, converging on a near-optimal global solution without requiring a centralized controller with global knowledge.

3. Heuristic-based Alternating Iterative Optimization Method (HAIOM)

Since caching decisions influence delivery efficiency and vice versa, the HAIOM approach iteratively refines both strategies. The system alternates between optimizing the caching plan (given a fixed delivery strategy) and the delivery plan (given a fixed caching strategy) until energy consumption stabilizes.

Performance Evaluation

The framework was evaluated using numerical simulations in a single-cell environment with 20 MUs. Content popularity followed a Zipf distribution, and channel conditions were modeled with Rayleigh fading. Social interactions were quantified based on contact duration and data packet sizes.

Key findings include:

  • Near-Optimality: The HAIOM-BPBP method achieved energy efficiency comparable to an Exhaustive Search (ES) while maintaining significantly lower computational complexity.
  • Efficiency Gains: Increasing device caching capacity led to marked reductions in both average energy consumption and delivery latency.
  • The "Rebound" Effect: The study noted that an excessive number of CHs can actually increase energy consumption due to interference and the overhead of the BS pushing content to too many devices.
  • Cache Hit Ratio: The joint optimization approach consistently outperformed heuristic methods (e.g., popularity-only or SINR-only priority) in terms of cache hit ratios.

Conclusion

By integrating socially-aware helper selection with distributed belief propagation, the proposed HAIOM-BPBP framework provides a robust solution for energy-efficient content delivery in D2D-enabled networks. Future research will explore dynamic content update strategies to account for rapidly shifting popularity trends and mechanisms to handle user churn as helpers enter or leave the network.

Summary updated

Abstract

Wireless distributed cache system (WDCS) is a promising technique to offload traffic and improve energy efficiency in content-centric networks. In this paper, the content sharing problem is considered by minimizing the average energy consumption for content caching and delivering in device-to-device (D2D) enabled cellular networks. The problem is formulated as a joint optimization problem of multiple interactive variables, which is NP-hard. In order to solve this problem, we first decompose the original problem into three subproblems, which are content helper (CH) selection, content caching, and content delivering. Specifically, in CH selection, we propose a PageRank based distributed algorithm to estimate the centrality of mobile users (MUs) by jointly analyzing their social relationships and channel conditions. Based on the selected CHs, the content caching and content delivering are formulated into two weighted matching problems. Two belief propagation (BP) based distributed algorithms are then proposed by considering content popularity and channel states to solve the aforementioned problems. Furthermore, based on the correlation of the three subproblems, a heuristic-based alternating iterative optimization method (HAIOM) is proposed to improve the average energy consumption performance of the content delivering of the WDCS. Simulation results show that by jointly optimizing the CH selection, content caching, and content delivering, the proposed scheme is capable of reducing the average energy consumption and the average content delivery delay, as well as of increasing the caching hit ratio.
BibTeX
@inproceedings{chuan2020optimizing,
  author = {Jianbin Chuan and Bo Bai and Xuewei Wu and Hongming Zhang},
  title = {{Optimizing Content Placement and Delivery in Wireless Distributed Cache Systems Through Belief Propagation}},
  booktitle = {IEEE Access},
  year = {2020},
  url = {https://openreview.net/forum?id=vvnG1hWWmt1U},
}
Figure