Blockchain-based cooperative game bilateral matching architecture for shared storage
Lin Guanjie,Mingyuan Zeng,Zhiguang Shan,Kaishun Wu,Guan Wang,Kai Lei
DOI: https://doi.org/10.1016/j.future.2024.04.016
IF: 7.307
2024-04-18
Future Generation Computer Systems
Abstract:The development of IPFS (InterPlanetary File System) and blockchain-based distributed storage projects has brought new possibilities to the field of storage. This paper proposes a blockchain-based cooperative game bilateral matching architecture as a novel approach for shared storage networks. In traditional competitive (non-cooperative) game models, the allocation of storage resources is centered around pricing, leading to a scenario where node providers often engage in price competition to obtain greater rewards, resulting in an imbalance in resource allocation for both buyers and sellers. In contrast, a distributed storage model based on cooperative game theory can better facilitate cooperation and resource sharing among node providers. This paper designs a storage resource allocation algorithm based on the stable marriage matching algorithm, demonstrating the stability of this algorithm as a matching solution. The paper also analyzes the differences between cooperative and non-cooperative game models in the market, and explores an equilibrium pricing mechanism guided by supply and demand. Furthermore, the paper introduces a trading mechanism for storage resources, including publication standards and matching schemes, ensuring efficient and trustworthy interaction between storage suppliers and demanders in a decentralized network centered around storage resources, thus enabling the circulation of the value of storage resources. A prototype of a blockchain-based shared storage trading system is implemented in this paper, utilizing bilateral matching for tradings. System evaluation and experimental testing are conducted, with results showing that the average utility value of the matching trading mechanism proposed in this paper outperforms the Double Auction-based matching model under any Poisson distribution ( λ =0.1, 0.2, 0.3, ..., 0.9) conditions set in the experiments. Additionally, compared to the traditional approach of directly storing complete data content on the chain, the design proposed in this paper effectively reduces on-chain storage consumption by approximately 27.06%.
computer science, theory & methods