Two-Level Transmission Scheme for Cache-Enabled Fog Radio Access Networks

Shiwen He, Chenhao Qi, Yongming Huang, Qi Hou, Arumugam Nallanathan

TL;DR

This paper proposes a "two-level" transmission strategy for Fog Radio Access Networks (F-RANs) that leverages edge caching to slash latency and reduce fronthaul congestion. By splitting delivery into a local "cache-level" phase and a joint "network-level" phase, the scheme maximizes data delivery rates while allowing edge nodes to operate semi-autonomously using a novel decentralized optimization algorithm.

What Problem Does This Paper Address?

As we push toward 5G and beyond, the demand for ultra-low latency (sub-10ms) and massive data rates is hitting a physical bottleneck: the fronthaul link. In traditional Cloud Radio Access Networks (C-RANs), a central "brain" (the Baseband Unit or BBU) does all the heavy lifting and sends signals to simple Remote Radio Heads (RRHs). This creates a massive traffic jam on the links connecting them and introduces significant processing delays.

Fog Radio Access Networks (F-RANs) try to solve this by adding "fog" capabilities—storage and processing—to the edge. These "enhanced" RRHs (eRRHs) can cache popular files locally. However, the challenge lies in coordination: How do we optimize the transmission of both cached and uncached files to maximize the delivery rate without overwhelming the fronthaul or requiring constant, high-latency signaling between edge nodes?

Methods

The authors introduce a Two-level Transmission Scheme:

  1. Cache-level Transmission: While the BBU is busy processing data or transferring files over the fronthaul, the eRRHs immediately start transmitting requested files that they already have in their local caches. This "hides" the BBU's processing delay.
  2. Network-level Transmission: Once the BBU is ready, all eRRHs work together via joint transmission to deliver the remaining requested files to users.

The Technical Innovations: * Centralized vs. Decentralized Optimization: For the cache-level phase, the authors developed a centralized algorithm (optimal but high overhead) and a decentralized algorithm. * SILNR (Signal-to-Interference-Leakage-plus-Noise Ratio): To make the decentralized approach work, they introduced the concept of SILNR. This allows an eRRH to optimize its beamforming by considering not just the interference it receives, but the interference it "leaks" to other users. This eliminates the need for eRRHs to talk to each other during the optimization process. * Convex Approximation: Since maximizing delivery rates involves non-convex "Signal-to-Interference-plus-Noise Ratio" (SINR) constraints, the authors used mathematical relaxations and successive convex approximation to ensure the algorithms converge to stable, high-performance solutions (KKT points).

Datasets

As a theoretical communication systems paper, this work does not use a standard "big data" dataset. Instead, it uses a Monte Carlo simulation environment based on standard wireless channel models:

  • Environment: A circular cell with a 500m radius.
  • Nodes: 3 eRRHs serving 5 single-antenna users.
  • Channel Model: Rayleigh fading with path loss (modeled with a power decay factor $\alpha=3$).
  • Library: A library of 10 files, where eRRHs pre-store files based on popularity and cache capacity.

Key Insights and Ideas

  • Latency as an Opportunity: Most papers treat BBU processing time as a penalty. This paper treats it as a "free" transmission window for cached content, effectively overlapping two different tasks to improve throughput.
  • The Power of Autonomy: The decentralized algorithm is the "aha moment." By using SILNR, an edge node can be a "good neighbor" by minimizing its leakage to others without needing a central controller to tell it exactly what to do.
  • Fronthaul-Awareness: Unlike simpler models, this scheme explicitly accounts for quantization noise—the distortion created when the BBU compresses signals to fit them through the limited-capacity fronthaul pipes.

Key Results

  • Near-Optimal Decentralization: The decentralized cache-level algorithm achieved over 97% of the performance of the centralized version. This is a huge win, as it provides nearly the same speed with significantly less signaling overhead.
  • TDR Gains: The Total Delivery Rate (TDR) increases significantly as the fronthaul capacity and file sizes increase, but eventually saturates as the system becomes "interference-limited" rather than "bandwidth-limited."
  • Superiority over Traditional Methods: The two-level scheme consistently outperformed traditional Cloud-Joint Transmission (TC-CJTM) because it makes better use of the time domain—starting the "local" party before the "cloud" data even arrives.

Limitations and Future Work

  • File Partitioning: The current model assumes files are either fully cached or not cached at all. In reality, "coded caching" (storing fragments of files) could offer even better performance.
  • Single-File Requests: The math assumes each user wants exactly one file. Future work could explore multi-file requests or spatial multiplexing for users with multiple antennas.
  • Multicast Scenarios: The paper focuses on unicast (one file to one user). If multiple users want the same viral video, a multicast approach would be more efficient.
  • Compression Algorithms: The authors noted that using more advanced signal compression on the fronthaul could further boost the "Network-level" phase performance.

Summary updated

Abstract

In this paper, we investigate the downlink transmission for cache-enabled fog radio access networks aiming at maximizing the delivery rate under the constraints of fronthaul capacity, maximum transmit power, and size of files. To reduce the delivery latency and the burden on fronthaul links and make full use of the local cache and baseband signal processing capabilities of enhanced remote radio heads (eRRHs), a two-level transmission scheme including cache-level and network-level transmission is proposed. In cache-level transmission, only requested files cached at the local cache are transmitted to the corresponding users. The duration of cache-level transmission is the delay caused by the transfer between the baseband unit (BBU) and eRRHs as well as the signal processing at the BBU. The remaining requested files are jointly transmitted to the corresponding users at network-level transmission. For cache-level transmission, a centralized optimization algorithm is first presented and then a decentralized optimization algorithm is provided to avoid the exchange of signaling among eRRHs. Meanwhile, another centralized optimization algorithm is presented to tackle the optimization problem for network-level transmission. All presented algorithms are proved to converge to the Karush-Kuhn-Tucker solutions of the problems. Numerical results are provided to validate the effectiveness of the proposed transmission scheme as well as evaluating the system performance.
BibTeX
@article{he2019two,
  author = {Shiwen He and Chenhao Qi and Yongming Huang and Qi Hou and Arumugam Nallanathan},
  title = {{Two-Level Transmission Scheme for Cache-Enabled Fog Radio Access Networks}},
  journal = {IEEE Trans. Communications},
  year = {2019},
  url = {https://openreview.net/forum?id=WLv72S3wzZ9j},
}

Loading related content...

Figure