| As an essential aspect of the Industrial Internet of Things(IIoT),the considerable amount of data can be utilized to achieve the generalized interconnection of all elements,including industrial personnel,machinery,and devices.However,there are some challenges that remain to be addressed,such as data trustworthy transmission,privacy protection,and collaborative sharing in the current multi-agent and low trust environment.These issues significantly hinder the advancement of industrial and value chains.Blockchain,serving as the fundamental technology that underpins data sharing in the IIoT,possesses many exceptional characteristics that distinguish it from other technologies.Trustlessness,transparency,and decentralization are some of the most powerful attributes.These promising features can promote the implementation of secure and transparent distributed systems,establish a trusted data management system in an untrusted competitive environment,and achieve cross domain data collaboration.Nevertheless,when applied to practical IIoT,the issue of scalability becomes a prominent concern.Due to the significant discrepancy between the performance indicators such as consensus speed and storage capacity of most blockchain technologies and the low latency application of massive terminal scenarios in the IIoT,it is difficult to meet the performance requirements of the industrial manufacturing field.Furthermore,there are notable disparities in the capabilities of industrial internet terminals and uneven resource allocation,which pose challenges in enabling a multitude of resource-constrained edge nodes to support the computing capabilities of consensus nodes in blockchain networks.Therefore,it is urgent to carry out research on the scalability of blockchain and edge computing offloading in the IIoT.This article aims to address the challenges associated with integrating blockchain and IIoT while towards the goal of low latency and high reliability.Encompassing network interconnection,data interoperability,security and trustworthiness,the focus centers on the pressing requirements for blockchain to empower the IIoT by enabling cross-chain data trusted sharing,cost-effective storage solutions,and scalable network transmission.Through the integration of edge computing,cryptography and other relevant technologies,and taking blockchain performance scalability of throughput,storage and network research as the starting point,an endogenous and reliable industrial scalable blockchain model is constructed to solve the problem of trust transmission and sharing of industrial multi-domain data,while also optimize the allocation of computing resources in the IIoT.1)To tackle the challenges arising from the performance limitations of conventional single-chain blockchain technology,particularly concerning consensus speed,which fail to meet the requirement for low-latency applications in the massive terminal environments of the IIoT,a scalable multi-edgechain and cloudlet chain structure suitable for IIoT is constructed.This architectural innovation greatly improves the difficulty of block tampering and breaks through existing single chain performance bottlenecks.Within this devised multi-chain structure,a Double-Level improved Practical Byzantine Fault Tolerance(DLPBFT)consensus algorithm is proposed.The first layer of DLPBFT consensus is independently completed by each edgechain and can be carried out concurrently.Subsequently,the second-level of DLPBFT consensus is achieved by the main nodes of each edge chain participating in the first round of consensus,and the key calculation result information is stored in the cloud chain to improve consensus efficiency.Compared with existing consensus methods,the throughput is increased by 27.7% and the block confirmation delay is reduced by 14.3%.In addition,to address the issue of data exchange between heterogeneous blockchains,a data sharing mechanism based on relay chain/side chain cross chain technology is designed and standardized for multi chain interaction models and protocols,achieving reliable data sharing between edge heterogeneous blockchains in the IIoT.2)To address the issue of the decline in consensus performance caused by the inability of the aforementioned consensus algorithm to adapt to node dynamic changes,as well as the high bandwidth of existing collaborative on chain storage methods for repairing failed node data,a scalable model for Dynamic Multi-domain collaborative Fractional Repetition(DMFR)code on chain storage is proposed.Firstly,the proposal involves designing a consensus mechanism for scalable dynamic multi-domain on chain storage to address the issue of decreased consensus performance caused by dynamic changes in the of number nodes and network connections within the current alliance chain’s admission and exit processes.It also mitigates the complexity of consensus communication before block storage.Secondly,based on the dynamically partitioned domains,a time-varying master node selection mechanism is introduced to tackle the inadequacy of existing practical byzantine fault-tolerant protocol master node selection methods in adapting to dynamic networks.Lastly,in view of the problems that the construction parameters of some existing duplicate codes are fixed,cannot be adjusted in time,and the repetition degree of isomorphic fractional repetition codes is increasing continuously,which cannot meet the needs of industrial data storage,a shadow-based heterogeneous fractional repetition code construction method is adopted to store block fragments in dynamically formed domains.Compared with existing FR code construction methods,the communication cost of recovering block data from failed nodes is reduced by an average of 29%.3)Focusing on the privacy and security issues in the process of cross domain data sharing and transmission,a scalable block compression network transmission model based on lattice proxy re-encryption is constructed to enhance the efficiency of secure data sharing among resource-constrained devices.Initially,to combat collusion attacks within proxy re-encryption,a novel blockchain-based distributed proxy framework is introduced to bolster system robustness.Concurrently,a lattice-oriented blockchain identity threshold proxy re-encryption is constructed to solve the complex certificate management problems in the current IIoT and resist quantum attacks.To ensure the verification of re-encrypted key share,the Merkel tree of blockchain is utilized to store key share on the chain.Secondly,a revocation operation for the re-encryption key is added to guarantee the security of the re-encryption key.Additionally,in the process of blockchain nodes serving as distributed agents for on-chain broadcasting of reencrypted key shares,considering factors such as node network latency,a node trust model is constructed and then the Kademlia broadcasting algorithm is optimized.Lastly,the Chinese remainder theorem is employed to compress block data in transmission.Compared to traditional broadcast mechanisms,the proposed approach exhibits an average 23.4% increase in effective transmission rates and a 19.8% reduction in broadcast time.These improvements significantly heighten network transmission scalability.4)Based on the built scalable blockchain model,a two-stage Stackelberg game method that incorporates both risk factors and offloading preferences is put forward.Determining the optimal amount of offloading computation and reward for undertaking computational tasks based on gradient descent algorithm,which effectively solve the real-time problem of edge computing tasks processing for edge devices with limited computing resources in the IIoT.It also majorizes the assignment of edge computing resources,thereby providing strong support for enhancing the efficiency of data sharing in the IIoT.Furthermore,the scalable blockchain model is applied to the offloading process of edge computing in dynamic networks.Given the mobility of nodes in the IIoT,there is a risk that edge devices in dynamic networks that respond to publishing computing tasks at a certain time may exit the network and make the remaining computing tasks impossible to complete.Therefore,considering the computational power of nodes,the responsive computing task is modeled as a dynamic stochastic game.To reduce computational complexity,the Time Varying Mean Field(TVMF)method is proposed for dynamic computation offloading optimization and obtain the optimal offloading control strategy under dynamic networks. |
PDF Full Text Request