Resilient backhaul network design using hybrid radio/free-space optical technology

Ahmed Douik, Hayssam Dahrouj, Tareq Y. Al-Naffouri, Mohamed-Slim Alouini

TL;DR

This paper proposes a cost-effective method for designing resilient mobile backhaul networks by combining traditional Optical Fiber (OF) with hybrid Radio Frequency/Free-Space Optical (RF/FSO) links. By formulating the network design as a graph theory problem—specifically a maximum weight clique problem—the authors demonstrate a way to guarantee high data rates and reliability through $K$ link-disjoint paths while significantly undercutting the deployment costs of fiber-only infrastructure.

What Problem Does This Paper Address?

As mobile data traffic continues its exponential climb, service providers face a "backhaul bottleneck." Connecting dense networks of small cells to the core network requires high capacity, but the traditional solutions have major drawbacks. Optical fiber (OF) offers massive speed but is prohibitively expensive and slow to deploy due to the need for physical trenching. Radio Frequency (RF) is easy to deploy but limited in bandwidth. Free-Space Optical (FSO) technology—which uses lasers to transmit data through the air—offers fiber-like speeds but is notoriously sensitive to weather conditions like fog and rain.

The authors tackle the challenge of designing a resilient backhaul: a network that remains functional even if specific links fail. They aim to minimize the total deployment cost while satisfying three strict constraints:

  1. Connectivity: Every node must be connected to every other node via at least $K$ paths that do not share any links (link-disjoint paths).
  2. Reliability: The network must meet a specific reliability threshold ($\alpha$), accounting for the fact that FSO links can fail in bad weather.
  3. Data Rate: Each node must meet a minimum target throughput ($D_t$).

Methods

The authors treat the backhaul design as a 0–1 mixed integer programming problem, which is computationally "heavy" (NP-hard) to solve directly. To make it tractable, they employ a two-step approximation strategy:

  • Step 1: The Optical Fiber Baseline: They first solve the problem assuming only optical fiber is used. They develop Algorithm 1, a clustering-based approach that iteratively builds a network with $K$ link-disjoint paths by connecting clusters using the cheapest available links.
  • Step 2: The Hybrid Reformulation: They then introduce hybrid RF/FSO links into the mix. Based on the observation that fiber costs scale with distance (digging) while hybrid link costs are dominated by fixed transceiver prices, they assume fiber is cheaper for short distances and hybrid links are cheaper for long ones.
  • Graph Theory Mapping: The problem is reformulated as a Maximum Weight Clique problem in a "planning graph." In this graph, vertices represent possible combinations of connections for a base station that satisfy reliability and rate constraints. Finding the "heaviest" clique (a subset of mutually adjacent vertices) corresponds to finding the lowest-cost valid network configuration.

Datasets

The paper relies on numerical simulations rather than a static external dataset. The parameters used include:

  • Environment: A $5 \text{ km} \times 5 \text{ km}$ square area with base stations placed randomly.
  • Cost Models: Optical fiber cost is set at $\$13.5/\text{meter}$, while hybrid RF/FSO transceivers are estimated at $\$20,000$ per link.
  • Reliability Model: Reliability for hybrid links is modeled as a function of distance, staying stable up to $2 \text{ km}$ before decaying exponentially.
  • Benchmarks: The proposed heuristic is compared against "Optimal Planning" (brute-force search) and "Optical Fiber Planning" (fiber-only).

Key Insights and Ideas

  • Redundancy as a Design Requirement: By enforcing $K$ link-disjoint paths, the authors move beyond simple connectivity to "resilience." This ensures that the failure of a single link (or even $K-1$ links) doesn't isolate a base station.
  • The Cost-Distance Tradeoff: The core "aha moment" is leveraging the different cost structures of OF and RF/FSO. Because fiber requires physical labor per meter, hybrid links become the "long-haul" heroes of the network, while fiber handles the "short-hop" connections.
  • Complexity Reduction: By mapping the problem to a maximum weight clique, the authors can use existing, highly efficient algorithms to find solutions that are nearly as good as the theoretical optimum but require a fraction of the computing power.

Key Results

  • Close-to-Optimal Performance: Simulations show that the proposed heuristic results in a total network cost that is nearly identical to the brute-force optimal solution, even as the number of base stations ($M$) increases.
  • Hybrid Efficiency: As the requirement for resilience ($K$) increases, the percentage of hybrid RF/FSO links used in the network grows. This is because hybrid links are more cost-effective at providing the "extra" paths needed for high redundancy.
  • Scalability: The total cost of the network scales much more gracefully with the number of nodes when using the hybrid approach compared to a pure fiber approach. For example, in a 7-node network with $K=4$ paths, the hybrid solution is significantly cheaper than the fiber-only baseline.

Limitations and Future Work

  • Line-of-Sight (LOS) Assumption: The model assumes all base stations have a clear line-of-sight to one another. In dense urban environments, buildings often block these paths.
  • Simplified Weather Modeling: While the paper accounts for reliability, it uses a simplified distance-based decay model. Real-world FSO performance varies wildly based on local micro-climates (e.g., specific fog pockets).
  • Future Directions: The authors suggest that future work could incorporate "non-line-of-sight" (NLOS) solutions, such as using mirrors to bounce FSO signals around obstacles, further increasing the flexibility of the hybrid design.

Summary updated

Abstract

The radio-frequency (RF) technology is a scalable solution for the backhaul planning. However, its performance is limited in terms of data rate and latency. Free Space Optical (FSO) backhaul, on the other hand, offers a higher data rate but is sensitive to weather conditions. To combine the advantages of RF and FSO backhauls, this paper proposes a cost-efficient backhaul network using the hybrid RF/FSO technology. To ensure a resilient backhaul, the paper imposes a given degree of redundancy by connecting each node through K link-disjoint paths so as to cope with potential link failures. Hence, the network planning problem considered in this paper is the one of minimizing the total deployment cost by choosing the appropriate link type, i.e., either hybrid RF/FSO or optical fiber (OF), between each couple of base-stations while guaranteeing K link-disjoint connections, a data rate target, and a reliability threshold. The paper solves the problem using graph theory techniques. It reformulates the problem as a maximum weight clique problem in the planning graph, under a specified realistic assumption about the cost of OF and hybrid RF/FSO links. Simulation results show the cost of the different planning and suggest that the proposed heuristic solution has a close-to-optimal performance for a significant gain in computation complexity.
BibTeX
@inproceedings{douik2016resilient,
  author = {Ahmed Douik and Hayssam Dahrouj and Tareq Y. Al-Naffouri and Mohamed-Slim Alouini},
  title = {{Resilient backhaul network design using hybrid radio/free-space optical technology}},
  booktitle = {ICC},
  year = {2016},
  url = {https://openreview.net/forum?id=LawOhHnCt2Z},
}
Figure