Os-elm based storage strategy for efficient query in blockchain database

Blockchain is widely used in supply chain systems due to its decentralization and security characteristics. However, with the outbreak of COVID-19, the supply chain systems need to have a more efficient query speed to meet the demand of origin-tracing. This paper proposes a hot block storage strategy to reduce the response time of data queries for blockchain-based supply chain systems. First, the architecture of the hot block storage strategy based on Online Sequential Extreme Learning Machine (OS-ELM) is designed, which includes the sharding-based blockchain module, the feature extraction module and the classifier module. Secondly, the fixed feature of the blocks, the node performance and other three features are considered to comprehensively evaluate the hot blocks for blockchain nodes. Third, the update algorithm for hot blocks is presented to ensure the persistence of efficient query. The experimental results show that the new strategy effectively improves the data query speed. Meanwhile, compared with the existing machine learning methods, the proposed OS-ELM-based strategy can significantly reduce the training time under the premise of ensuring the accuracy of classification when new training data is added. The OS-ELM-based storage strategy can achieve quick origin-tracing for blockchain-based supply chains and is suitable for the systems with rapidly increasing data.