Research On Computing Offloading And File Storage Strategies For Blockchain Networks

In recent years,the rise of blockchain technology has aroused widespread concern in the academic community.As a distributed computing and storage method,nodes in the blockchain network have to undertake heavy tasks,including consensus computing represented by mining and storage of massive data.However,some mobile miners are unable to complete mining tasks due to their insufficient computing capabilities.With the advent of Mobile Edge Computing(MEC),by combining the blockchain and wireless networks,the problem of miners' insufficient capacity can be effectively solved through offloading mining tasks to edge computing nodes.Multiple factors need to be measured to find the best offloading solution to optimize system performance.On the other hand,since the size of each block in the blockchain is generally only a few megabits,large files cannot be stored.The traditional strategies of blockchain store files on a designated cloud server and generate a hash value to ensure security.However,the schedule of the endorsement nodes and the storage nodes is not taken into account by those strategies during the storage unloading process,which planning will cause problems in terms of resources and security.In addition,the frequency of file usage is also not considered by the traditional strategies,which will cause backhaul link pressure for frequent downloading of files from cloud servers.Based on this,in order to effectively improving the offload energy efficiency and large file storage efficiency of the wireless blockchain network,this paper deeply studies the energy-efficient task offload strategy and the file storage strategy based on the Fabric platform in the wireless blockchain network.The specific research content and main contributions are as follows:First,for the offloading of blockchain network tasks,two offloading strategies based on energy efficiency optimization are proposed.Through modeling,the mining task is divided into two parts: local computing and offload processing.The fair energy efficiency is defined as the ratio of user rate to power consumption.The network resource optimization problem under both the maximum-minimum rate criterion and the maximum system energy efficiency criterion are studied,respectively.TheDinkelbach algorithm and SCALE(Successive Convex Approximation for Low Complexity)algorithm are applied to solve the optimization problem,and the proposed schemes are compared with the two benchmarks,i.e.,system power consumption minimization method and system processing rate maximization method.The simulation results verify that the offloading scheme proposed in this paper can obtain the best energy efficiency of the network.At the same time,it shows that there is an optimal tradeoff between power consumption and speed,which has guiding significance for practical network design.For the blockchain file storage part,a large file storage system based on the Fabric platform was constructed.According to the storage rules of Fabric,endorsement(that is,verification of transactions)is required before storage.1)Aiming at the endorsement process,a dynamic endorsement node selection strategy is designed,by using the sum of upload delay,queuing delay,and processing delay as the communication node's score,and selecting the optimal endorsement node based on the score can effectively improve the verification speed.2)For the storage process,a distributed file classification storage scheme based on the blockchain is designed:scoring files by delay,and then planning corresponding storage nodes for the files based on the ratings,which can enhance storage security and resource usage efficiency.3)With regard to the file reading process,by intelligently designing smart contract monitoring files,effectively monitoring the number of calls of each file within a specified time,then dividing the files into two categories based on the frequency,By storing frequently used files near the user,which allows users to download directly from nearby storage devices without using the cloud backhaul link.The experiments show that the large-scale file storage system based on the Fabric platform designed in this paper can not only ensure the efficiency of endorsement nodes with alleviate the pressure on the clients' selection,but also quickly recover damaged files and solve the problem of data loss.The system also can speed up file reading speed,alleviate backhaul link pressure,and reduce problems such as line congestion.