| Smart contracts,as a kind of code script deployed on the blockchain,can automatically complete a predefined business or computation for the user when called by a transaction.However,as blockchain application scenarios become diverse,smart contracts are becoming increasingly complex,and the amount of computation contained in smart contracts has seen a significant increase.Therefore,computationally intensive contract transactions are used to refer to such invocation transactions of smart contracts where the validation process consumes a lot of computational resources.How to achieve fast verification of computation-intensive contract transactions by blockchain has become an important research for blockchain to break through the performance bottleneck and achieve multi-scenario implementation.Among the existing studies,the strategy of using parallel execution to shorten the transaction verification time has been shown to be an effective means to improve the throughput performance of blockchains.However,most of these schemes are proposed for simple transactions and do not consider the special needs of computationally intensive contract transactions,and do not generate a systematic discussion between parallel techniques and computationally intensive contract transactions to guide the combination of the two.In addition,blockchain has limited ability to support computationally intensive contract transactions.On the one hand,smart contracts have poor multicore adaptability and suffer from replay indeterminacy when being executed in parallel.On the other hand,computationally intensive contract transactions lack a verifier dilemma solution with good on-chain and off-chain data consistency.To this end,this thesis makes efforts in the following aspects.1.For the lack of theoretical guidance on the application of existing parallel schemes to computationally intensive contract transactions,the thesis proposes two parallel classifications for computationally intensive contract transactions.Using a typical computation-intensive contract case,the thesis analyzes the special requirements of computation-intensive contract transactions,broadly classifies the existing parallel techniques into two categories,and discusses the solutions in both categories of parallelism in several dimensions,explores the possibility of integration and combination of existing solutions from the perspective of computation-intensive contract transactions,and analyzes the two shortcomings of computational intensive contract transactions in the parallelization process.2.To address the problem of replay uncertainty when smart contracts are executed in parallel by multiple cores,the thesis proposes parallel programming contracts that support replay determination.This includes providing a smart contract model that can allow multiple computing cores to verify a contract transaction simultaneously,improving the verification efficiency of computationally intensive contract transactions.At the same time,the model can solve the problem of disorderly competition for resources that exists in parallel tasks and guarantee the replay determination of parallel results.3.To address the verifier’s dilemma of computationally intensive contract transactions,the thesis proposes an off-chain computation scheme based on an on-chain challenge protocol.The scheme can allow computation-intensive transactions to not have to be repeatedly computed in the consensus process,while the transaction results can still get correct consensus,and the computation process also has the grouping parallelism feature.This model solves the problem of short-term inconsistency of on-chain and off-chain data that existed in previous similar models,and can effectively reduce invalid transactions while significantly improving blockchain throughput.The on-chain challenge protocol also contains the corresponding incentive model and resource over-delivery strategy,which further guarantees the reliability and availability of the on-chain challenge protocol. |
PDF Full Text Request