Decentralized storage networks increasingly rely on blockchain-based verification to ensure data integrity without centralized control; however, proof-intensive workloads introduce significant latency and on-chain cost overhead. This paper presents a lifecycle-based comparative analysis of major zero-knowledge proof (ZKP) models used in decentralized storage, focusing on zk-SNARK frameworks and transparent zk-STARK constructions. A multi-layer evaluation framework is introduced, aligning performance analysis with the core stages of the proof lifecycle: generation, aggregation, and on-chain verification. Building on this analysis, the paper proposes a hybrid architecture that combines parallel STARK-based proof generation with recursive SNARK-based compression, reducing on-chain verification complexity to near-constant. A Filecoin-inspired case study, supported by a quasi-empirical performance model, demonstrates that the proposed hybrid approach significantly reduces verification latency and data overhead while mitigating the linear growth of verification costs. The results indicate that hybrid ZKP architectures offer a scalable and economically viable solution for decentralized storage systems and large-scale blockchain networks................. Keywords: .............Zero-knowledge proofs, zk-SNARKs, zk-STARKs, recursive aggregation, decentralized storage, verifiable cryptography, scalability, gas cost.