Ilmu Komputer & AI editorial
Split-Aware Function Placement with Availability Guarantees and Optical Provisioning in vRANs
The core problem
The rapid evolution of beyond-5G and emerging 6G networks demands flexible, reliable, and cost-efficient virtualized Radio Access Network (vRAN) architectures. These must support heterogeneous services including enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communication (URLLC), and Massive Machine-Type Communication (mMTC). Future disaggregated RAN systems are expected to rely heavily on network slicing, functional split flexibility, and optical x-haul infrastructures to meet stringent performance, scalability, and availability requirements.
This paper addresses the challenge of jointly optimizing Virtual Network Function (VNF) placement and optical resource allocation in such environments. The goal is to maximize mobile network operators' profit while satisfying latency, processing, bandwidth, and availability constraints. The authors propose an integrated framework that is reliable, slice-aware, and functional split-aware, and they formulate the problem as an Integer Linear Programming (ILP) model with two backup variants: unshared and shared. To overcome ILP complexity, they develop a heuristic algorithm and a Genetic Algorithm (GA)-based metaheuristic that yie
Innovation
Extensive evaluations were conducted on topologies up to 128 nodes. Key findings include:
- **Profit**: Shared backup variants yield up to **18% higher profit** compared to unshared backups.
- **CPU usage**: Shared backups maintain up to **5–10% lower normalized CPU usage** than unshared counterparts.
- **Scalability**: The heuristic and GA-based metaheuristic produce near-optimal solutions in real time, making them suitable for dynamic vRAN environments.
The results demonstrate that sharing backup resources significantly improves cost efficiency without compromising availability guarantees. The GA consistently outperforms the heuristic in terms of profit, especially in larger topologies, while both scale well beyond the capabilities of exact ILP solvers.
Why it matters
The paper highlights the importance of joint optimization of function placement and optical provisioning in disaggregated vRANs. By considering functional splits and slice awareness, the framework ensures that diverse service requirements (eMBB, URLLC, mMTC) are met. The shared backup scheme emerges as a key enabler for cost-effective reliability, as it reduces redundant resource allocation while still satisfying availability constraints.
The ILP model provides a solid theoretical foundation, but its complexity necessitates heuristic and metaheuristic approaches for practical deployment. The GA-based method offers a good balance between solution quality and computational time, making it attractive for real-time orchestration.
Future work could extend the framework to handle dynamic slice arrivals and departures, integrate machine learning for predictive placement, and consider energy efficiency as an additional objective. The taxonomy candidates (Architecture, Cybersecurity, Network, Cryptography) suggest potential cross-domain applications, particularly in secure network slicing and resilient infrastructure design.
Who should read this
Opening member content…