| In recent years,the housing rental industry has been booming.However,there are still many problems in various online rental platforms,such as the proliferation of fake housing resources,the mixed intermediary and the principal landlord,and the lack of trust between the two parties.To solve the above problems,this thesis designs and implements an online rental system based on micro-service and blockchain technology.It also improves the blockchain consensus algorithm.The system contains three sub-modules.The first is the user module for the landlord and the tenant,which includes user registration,message notification and personal information management.The second is the house module,which mainly includes the functions of releasing the house,searching the house,displaying the house and making an appointment to see the house.The third is the certification module based on Huawei cloud blockchain platform,which connects the core data of rental business.Through the tripartite authentication of the Public Security Bureau,the Housing Administration Bureau and the Credit Information Center to ensure the authenticity of housing sources and user information.All user information,house source data and lease contract involved in this system are stored in the form of IPFS hash on the blockchain smart contract to ensure the security and traceability of the data.The system split the services according to different business functions.This thesis designs the system using the micro service architecture with strong expansibility and low coupling degree.The system uses Dubbo as the distributed service framework.It employs Zoo Keeper as the registry to manage and schedule services.It also realizes the communication between services by means of remote call.To cope with high concurrent requests,the system adopts Redis to cache data and Rocket MQ message queue to process requests asynchronously.That reduces the pressure on My SQL database and improves the system throughput.The server of the system is implemented on Spring Boot and Spring MVC framework.After the development is completed,black-box test is used to verify the system function.In this thesis,a consensus algorithm based on Practical Byzantine Fault Tolerance(PBFT)is selected for the blockchain authentication platform.And a method to improve the scalability of the consensus algorithm is proposed.In a system with N nodes,the communication complexity of PBFT reaches O(N~2).Considering the existence of network delay,the efficiency of PBFT algorithm will decrease rapidly when the number of nodes reaches a certain number,which seriously restricts its scalability.To solve these problems,a consensus algorithm of PBFT based on Consistency Coordinator is proposed in this thesis.First,a master node is selected in the blockchain network to execute the consensus algorithm and communicate with the conformance coordinator.The coordinator then collects transactions from the transaction pool on each node.Transactions are then classified by equality,and only if some transactions are not equal will the coordinator ask the master node to perform a consensus algorithm on the failed transactions among them.Finally,the coordinator sorts all the agreed transactions in chronological order and asks the master node to generate a new block containing all the processed transactions.Experimental shows when the number of system nodes increases from 4 to 80,the transaction delay of the traditional PBFT algorithm increases 50.29 times.The transaction delay of the PBFT algorithm with the coordinator increases 34.12 times,which is only 67.8%of the former.The scalability of the system is significantly improved.This thesis uses the micro service architecture to realize the online rental platform.Besides,it improves the scalability of the PBFT consensus algorithm in the blockchain certification module,so that it can be applied in this system on a large-scale.At present,this system has been running stably in a domestic Internet rental company for nearly a year. |
PDF Full Text Request