Hyra: Scalable Byzantine-Resilient State Storage Engine with Hierarchical Erasure-Coding
Account-based blockchains maintain evolving state data, such as account balances, across all replicas to ensure consistency. However, the traditional full-replication model incurs significant storage overhead and limits scalability, especially as transaction throughput increases. While partitioning is a natural solution, applying it to blockchain state is challenging due to structural dependencies, verifiability requirements, and the presence of Byzantine faults. We present Hyra, a scalable state storage engine that enables fault-tolerant and efficient state partitioning. Hyra introduces: (i) locality-aware partitioning with embedded indexing to preserve access efficiency; (ii) hierarchical erasure coding with sub-chunk recovery to minimize decoding overhead; and (iii) a hybrid verification mechanism that combines vector commitments and Merkle trees for fine-grained integrity checking. Our design achieves provably optimal redundancy while maintaining constant per-replica storage overhead. Experiments show that Hyra reduces storage by up to 92.3% without compromising performance.
Authors
- Aoying Zhou (ORCID: https://orcid.org/0000-0002-4665-7302)
- Xiaodong Qi (ORCID: https://orcid.org/0000-0002-2899-8656)
- Yanqin Yang
- Cheqing Jin
- Qifeng Que
Institutions
- Nanyang Technological University (SG)
- East China Normal University (CN)
Publication Details
- Journal
- Proceedings of the ACM on Management of Data
- Published
- 2026-04-02
- DOI
- https://doi.org/10.1145/3786658
- Primary Topic
- Distributed systems and fault tolerance
- Type
- article
- Field-Weighted Citation Impact
- 0.00