Cost-Effective Availability-Aware Dimensioning for Cloud RAN

We study 6G Cloud Radio Access Network (C-RAN) dimensioning, treating availability as a key metric alongside throughput and latency. Given fixed RU-to-site mappings and user availability targets, we formulate a cost-minimization problem that selects redundancy and component configurations using general-purpose servers and C-RAN software. We introduce two deterministic optimizers that guarantee optimal solutions: a per-gNB optimizer that finds the lowest-cost feasible plan for each gNB (i.e., 5G/6G base station), and a site-level optimizer that jointly configures multiple gNBs to exploit pooling. Compared with a representative heuristic baseline, the per-gNB method achieves similar runtime and consistently reaches the global optimum, whereas the site-level method delivers the lowest total cost at higher computational complexity.