Lookup Table Designs for Real-Time RAFT Consensus Agreements on HAPS Platforms

High-Altitude Platform Station (HAPS) Mesh networks hold promise for delivering high-availability wide-area network services, enabling real-time RAFT in distributed systems. However, the dynamic nature of HAPS networks disrupts RAFT's critical performance metric - the 95th percentile commit time - which directly impacts service level agreements (SLAs). To address this, we tackle the challenge of dynamically selecting cost-effective Steiner Minimum Trees (SMTs) within HAPS Mesh networks to connect RAFT leaders and quorums while strictly adhering to SLA constraints. Traditional RAFT performance modeling methods fall short in handling these applicationspecific requirements under real-time operational demands. Our solution introduces an efficient lookup table (LUT) framework for rapid assessment of SMT quality, overcoming the latency of direct evaluation. Key innovations include leveraging path score monotonicity, boundary point analysis, and a 'striping' strategy to map solution spaces while minimizing computational overhead. Extensive simulations and experiments demonstrate the feasibility and effectiveness of our proposed method.