Data Security Based On Blockchain And Federated Learning

Nowadays,the interconnection and utilization of marine information has been attached great importance to by the world.In the past,marine Internet of Things(MIoT)technology has mostly used cloud computing architecture.However,with the advent of the era of big data,cloud computing gradually faces many challenges such as network delay.The birth of edge computing solves these problems.However,due to the lack of trust among participants and user concerns about privacy,a more reliable solution is needed.Some of current solutions use blockchain technology to address data security issues,some use federated learning technology to address privacy issues,but these approaches do not incorporate the special environment of the ocean and do not consider the security of the nodes.This paper proposes a method for secure data sharing in the MIoT under the edge computing framework based on federated learning and blockchain technology.Federated learning combines its special distributed architecture with the edge computing architecture of the MIoT to ensure node privacy.Blockchain serves as a decentralized method for storing and managing federated learning against workers in order to achieve immutability and security.This paper introduces the concepts of quality and reputation as metrics for federated learning workers selection.This prevents the possibility of an edge node uploading unreliable data that could lead to spoofing in federated learning tasks.At the same time,this paper designs a proof of quality consensus mechanism(PoQ)and applies it to the blockchain,so that the edge nodes recorded in the blockchain are of higher quality and the overall model effect is improved.In addition,the marine environment model is also constructed in this paper,and the analysis based on this model makes the method proposed in this paper more suitable for the marine environment.The numerical results obtained by the simulation experiments clearly show that the proposed scheme can significantly improve the learning accuracy under the premise of ensuring safety and reliability in the marine environment.Finally,this paper designs a data security sharing scheme suitable for the MIoT.In this paper,the edge computing framework of the MIoT has more efficient data processing and more secure data protection capabilities,and this paper proves the effectiveness of the scheme through experiments.