Research On Blockchain System With Trustworthy Oracle And Lightweight Storage For Industrial Internet Of Things

At present,it has become an important trend to promote the application of blockchain in Industrial Internet of Things(IIo T).Blockchain can provide industrial sectors with a trusted data-sharing and computing environment,reducing the costs required by infrastructure construction.However,the deterministic model of blockchain smart contracts limits their ability to proactively interact with other systems and handle those computing tasks involving real-time external data.Meanwhile,massive on-chain industrial data bring considerable storage pressure,making it difficult for IIo T nodes with limited storage resources to participate in blockchains.Decentralized oracles can offer a reliable data acquisition mechanism for blockchain systems,allowing smart contracts to request data from external systems during their execution process.Among them,the ones based on Verifiable Random Function(VRF)can better adapt to the IIo T environment due to high security and node resource utilization.But,their complete randomness can undermine the quality of service and the willingness of nodes to continuously serve,thus affecting the quality and timeliness of feedback data.Besides,to reduce the storage requirement on a single node,the existing lightweight storage schemes usually require each node to maintain only a small part of blockchain data according to specific rules,under the premise of ensuring global data integrity.While executing smart contracts,a blockchain node might have to frequently access data in non-local ledgers,where the data requesting delay will lead to the limitation of performance.Therefore,this thesis focuses on improving the quality of oracle services for IIo T and the execution efficiency of smart contracts on storage-constrained IIo T nodes.The main contributions are summarized below.(1)Firstly,a distributed oracle scheme for IIo T is proposed.On the one hand,a VRFbased oracle node selection algorithm is designed to improve the quality of selected oracle nodes,where the reputations of nodes are considered within limits while keeping sufficient randomness and avoiding Matthew effect.Moreover,oracle nodes are enabled to keep anonymous before they publish their feedback,reducing the probability of being attacked.On the other hand,a data-filtering algorithm based on sliding window is introduced to optimize the time distribution of feedback data,thus improving the consistency.Aiming at off-chain aggregation,an implementation of this algorithm is presented,with Raft as the underlying consensus algorithm.The experiment results show that,with part of malicious nodes,the proposed oracle scheme can effectively improve the accuracy and consistency of feedback data and provide a shorter response time.(2)Secondly,a lightweight blockchain data storage scheme is proposed.On the one hand,according to the characteristics of IIo T,a general ledger structure,network architecture,and transaction flow are respectively defined.At the storage layer,blockchain nodes are allowed to use their unoccupied storage space to selectively cache another part of blockchain data in addition to their responsible part.At the network layer,the communication costs of data dissemination are optimized.On the other hand,a cache agent based on Deep Reinforcement Learning(DRL)is formulated,which enables storage-constrained blockchain nodes to decide which part of blockchain data to cache according to the access history of state data and smart contracts,maximizing the execution efficiency of contracts.An implementation based on Deep Q-learning is presented.The experiment results show that the proposed scheme can effectively reduce data dissemination costs in large-scale networks and data-requesting delays on storage-constrained nodes.In conclusion,in the aspect of data acquisition,the proposed oracle scheme can balance the service quality and security of decentralized oracles.In the aspect of data storage,the proposed blockchain storage scheme can achieve higher computing performance with limited storage space.Extensive experiments show that these schemes are effective in improving the service quality of blockchain oracle and reducing the execution efficiency of smart contract,and has practical application value.