Scholarlink

With the widespread adoption of blockchain technology, its inherent data storage bottleneck has become increasingly prominent. Traditional blockchain systems typically rely on full replication storage strategies, resulting in high storage overhead and poor scalability, which in turn limits their applicability in key domains such as the Internet of Things (IoT), finance, and supply chains. Although current approaches—such as off-chain storage and data sharding—have alleviated some of the storage pressure on blockchain systems, they still fall short of fundamentally overcoming the scalability constraints imposed by full replication. In contrast, erasure coding technologies, by leveraging data partitioning and redundant encoding mechanisms, provide a fundamental breakthrough to these limitations and have thus attract significant attention. However, existing erasure-code-based storage schemes continue to face challenges such as high encoding computational complexity, significant communication overhead, and difficulties in dynamic node adjustment. To address these challenges, this paper proposes HRE-Store, a blockchain storage scheme based on a hierarchical storage architecture. This design employs a hierarchical storage framework and partitioned encoding mechanism to decouple global encoding tasks, thereby reducing encoding complexity. A lightweight indexing mechanism is constructed to minimize communication overhead, and a minimum threshold mechanism for nodes is introduced to simplify the process of dynamic node adjustment. Theoretical analysis and experimental results demonstrate that, in large-scale node environments, HRE-Store achieves approximately 64 \(\times \) and 3 \(\times \) improvements in computational efficiency compared to the classical BFT-Store and the latest PartitionChain coding scheme, respectively. In terms of read performance, latency is reduced by an average of approximately 90% and 80%, respectively, while in terms of system stability, data recovery time is reduced by approximately 87% and 55%, respectively.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

HRE-Store: A Hierarchical and Scalable Storage Architecture for Permissioned Blockchains

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yang Liu,
  • XiangYu Cui,
  • FangChao Tian,
  • Feng Wang,
  • Han Li,
  • Min Zhang

摘要

With the widespread adoption of blockchain technology, its inherent data storage bottleneck has become increasingly prominent. Traditional blockchain systems typically rely on full replication storage strategies, resulting in high storage overhead and poor scalability, which in turn limits their applicability in key domains such as the Internet of Things (IoT), finance, and supply chains. Although current approaches—such as off-chain storage and data sharding—have alleviated some of the storage pressure on blockchain systems, they still fall short of fundamentally overcoming the scalability constraints imposed by full replication. In contrast, erasure coding technologies, by leveraging data partitioning and redundant encoding mechanisms, provide a fundamental breakthrough to these limitations and have thus attract significant attention. However, existing erasure-code-based storage schemes continue to face challenges such as high encoding computational complexity, significant communication overhead, and difficulties in dynamic node adjustment. To address these challenges, this paper proposes HRE-Store, a blockchain storage scheme based on a hierarchical storage architecture. This design employs a hierarchical storage framework and partitioned encoding mechanism to decouple global encoding tasks, thereby reducing encoding complexity. A lightweight indexing mechanism is constructed to minimize communication overhead, and a minimum threshold mechanism for nodes is introduced to simplify the process of dynamic node adjustment. Theoretical analysis and experimental results demonstrate that, in large-scale node environments, HRE-Store achieves approximately 64 \(\times \) and 3 \(\times \) improvements in computational efficiency compared to the classical BFT-Store and the latest PartitionChain coding scheme, respectively. In terms of read performance, latency is reduced by an average of approximately 90% and 80%, respectively, while in terms of system stability, data recovery time is reduced by approximately 87% and 55%, respectively.