Batch Transaction Compression for Efficient Off-Chain Storage in Blockchain-Based IoT Systems

Abstract

With the increasing adoption of blockchain systems in financial services, supply-chain management, and Internet-of-Things (IoT) applications, the volume of transaction data stored both on-chain and off-chain has grown rapidly, imposing significant challenges on storage costs, synchronization latency, and resource-constrained devices. In IoT environments, edge devices typically possess limited storage and computational capacity, making long-term archival of transaction records particularly challenging and detrimental to system scalability. To address these challenges, this paper proposes an off-chain archival framework that combines Schema-aware transaction encoding with high-performance lossless compression. A Schema-aware transaction encoding framework (TXPACK) is designed to transform batches of transactions into representations that are more amenable to compression, followed by a second-stage compression using ZSTD. Cryptographic hashes of the compressed data are then committed to the blockchain to ensure data integrity and on-chain verifiability while preserving the existing blockchain security model. Experiments were conducted on the NEAR blockchain test network using real transaction traces. The proposed TXPACK+ZSTD approach was compared with widely adopted lossless compression algorithms, including GZIP, LZ4, and standalone ZSTD, in terms of compression ratio, compression time, decompression latency, and transaction confirmation delay. The results show that TXPACK+ZSTD achieves a maximum space-saving ratio of 98.91% under homogeneous transaction batches, outperforming standalone ZSTD (97.96%), GZIP (97.94%), LZ4 (94.84%), and Brotli (98.00%). Under heterogeneous transaction datasets, TXPACK+ZSTD still achieves 93.99% space saving, remaining superior to standalone ZSTD (91.24%), GZIP (92.18%), and LZ4 (79.52%), while maintaining comparable compression and decompression overhead. On-chain evaluations further indicate that submitting compressed payloads does not introduce noticeable increases in transaction confirmation time. In addition, a formal security analysis demonstrates that the integrity of archived off-chain data can be guaranteed under standard cryptographic hash assumptions and the immutability of the underlying blockchain. Overall, this study confirms that integrating Schema-aware encoding with modern compression techniques can significantly reduce off-chain storage requirements for IoT-oriented blockchain applications without altering consensus mechanisms. To the best of our knowledge, this work is the first to integrate schema-aware encoding with Zstandard compression in a complete off-chain archival pipeline with on-chain hash commitments on the NEAR blockchain.