DLBN: Group Storage Mechanism Based on Double-Layer Blockchain Network

Blockchain, which stores data in an appending form, cannot achieve the purpose of expanding the storage capacity by increasing the number of nodes. As the system runs, nodes will face problems of insufficient storage space. In the existing peer-to-peer (P2P) blockchain network model, all network nodes participate in data storage, and the generated blocks need to be verified among the network-wide nodes. This approach suffers from low system transaction throughput and data storage redundancy. In order to solve the above existing problems, this article proposes a block data storage model based on the double-layer blockchain network (DLBN), which improves the internal data composition structure of the blockchain. The DLBN contains two types of blockchain nodes, which form the storage and consensus layers of the system, respectively. The consensus layer is responsible for tasks, such as transaction sequencing, validation, and block packing, thus increasing the system transaction throughput. The nodes in the storage layer are divided into multiple storage units (SUs), and all nodes in the SU jointly maintain a copy of the complete blockchain, thereby reducing the storage pressure on the nodes. Based on the DLBN model, we design a reputation-based consensus mechanism, block storage allocation algorithm, and transaction query optimization algorithm, respectively. Through experimental verification and analysis, the storage model based on the DLBN can effectively improve the system transaction throughput and reduce the node storage capacity while ensuring system security.

I. Introduction

Blockchain distributed ledger [1] has changed the traditional centralized database storage with its features, such as decentralization and tamper-proofness. The nodes in the network can realize peer-to-peer (P2P) transactions without relying on third-party trusted platforms, thus circumventing the data security problems that exist in centralized institutions. The blockchain technology is rapidly developing in various network-related fields, such as the Industrial Internet of Things [2], the Internet of Vehicles [3], crowdsourcing [4], and the smart grid [5]. As the scope of blockchain applications expands, many challenges also follow [6], [7]. As a decentralized data management platform, blockchain avoids tampering with user data and ensures the reliability and accuracy of data [8]. However, due to the inherent distributed property of the blockchain system, i.e., each transaction has to be confirmed by nodes across the network, resulting in low transaction throughput of blockchain. Compared with traditional databases, blockchain is limited inefficiency and also consumes larger storage capacity. Therefore, enhancing the transaction throughput of blockchain systems and reducing the storage pressure on nodes is gradually becoming a research hotspot.