Improving query processing in blockchain systems by using a multi-level sharding mechanism

• Published in: The Journal of supercomputing Vol. 80; no. 10; pp. 15066 - 15096
• Main Authors: Matani, Alemeh, Sahafi, Amir, Broumandnia, Ali
• Format: Journal Article
• Language: English
• Published: New York Springer US 2024; Springer Nature B.V
• Subjects: Blockchain; Compilers; Computer Science; Data structures; Heterogeneity; Interpreters; Nodes; Privacy; Processor Architectures; Programming Languages; Queries; Query processing; Relational data bases; Sharding; Heterogeneity; Query processing; Consensus; Blockchain
• ISSN: 0920-8542; 1573-0484
• DOI: 10.1007/s11227-024-06037-5

With the distributed and decentralized nature of blockchain, and with its sequential data access, query processing emerges as a challenging issue in the blockchain systems. These features hinder efficient query processing and make it difficult to guarantee the validity and privacy-preserving of query results. Several solutions have been proposed to tackle the efficiency, reliability, and privacy challenges of query processing in blockchain systems. There has been rarely a comprehensive solution addressing all of these issues. In addition, the existing solutions often assume that the blockchain nodes are homogeneous in terms of their capabilities and available resources, while the blockchain nodes can have heterogeneous computational, communication, and storage resources, and can also contribute to the blockchain network in different manners. This work, considering the heterogeneity of network nodes, introduces a multi-level and score-based sharding solution for query processing where the nodes are organized into a hierarchical tree-like structure based on their score and store a proportion of transaction data in a DAG-based data structure resulting in an efficient query time. Additionally, the nodes reach a consensus over the query results from the bottom to the top of the hierarchical structure enabling reliable and fast query processing. The experiments conducted during the evaluation show that the efficiency of the proposed work is near that of relational databases in terms of query response time. It also provides a high validity rate taking advantage of its hierarchical consensus mechanism and preserves the privacy of query results using a delegation-based integration method where the final query result is integrated by the client’s representative.