An Efficient Blockchain-Based Data Storage and Query System for Large-Scale Data Sharing

In blockchain–IPFS-based systems, full nodes maintain complete ledger replicas, whereas light nodes retain only essential information such as block headers to reduce storage and computation overhead. Due to the absence of full data replicas, light nodes are unable to support full-data queries, which limits their applicability in practical financial data sharing scenarios. Moreover, conventional blockchain storage mechanisms rely on synchronous confirmation across multiple nodes, resulting in limited throughput and poor responsiveness under high-concurrency and burst-traffic conditions. To address these issues, this paper proposes a blockchain–IPFS-based storage and query scheme for banking credit data that integrates multi-level caching, non-blocking asynchronous processing, and a Cuckoo filter–based lightweight query mechanism. The proposed scheme enables light nodes to efficiently verify the existence of credit files and retrieve associated metadata without maintaining complete ledger replicas, while a coordinated caching–asynchronous architecture decouples user requests from on-chain and off-chain persistence operations to improve system throughput and robustness. A prototype system is implemented and evaluated under varying file sizes and concurrency levels. Experimental results show that, for files smaller than 100 MB, the proposed scheme reduces storage latency by approximately 35–99% and improves query response time by more than 95%, compared with conventional blockchain–IPFS-based solutions. In addition, download latency is reduced by 20–31% for small and medium-sized files. The results further confirm that the proposed approach effectively supports full-data queries for light nodes and demonstrates strong resilience under burst traffic conditions, indicating its practical feasibility for secure financial credit data sharing.