Research And Application Of Distributed E-commerce Storage System Based On IPFS

In recent years,e-commerce has occupied an irreplaceable position in daily life due to its convenience.With the development of e-commerce systems,users and data have also exploded.The construction of e-commerce storage systems has become a key issue.At present,domestic and foreign e-commerce website storage systems are mainly composed of relational databases such as MySQL and distributed cluster systems such as HBase.There are still many problems with such centralized cluster storage systems,such as data backup security and service downtime.The interstellar file protocol IPFS is a fully distributed file system that uses peer space to store data and provide services,which can greatly optimize the disadvantages of existing e-commerce storage systems.This paper studies how to use the advantages of IPFS to build a distributed e-commerce underlying storage system to improve the data security and service availability of e-commerce systems.Based on in-depth research on IPFS technology,e-commerce storage architecture,and full-text information search technology,this paper designs a storage system for distributed e-commerce based on IPFS,which includes storage scheme design and search mechanism design.First,by analyzing the underlying storage mechanism of IPFS,a file storage scheme for e-commerce data is designed,and an IPFS file high-availability backup algorithm based on the QoS and interest of IPFS nodes is proposed to solve the intermittent data unavailability in traditional IPFS networks.At the same time,research on the IPFS search system,and propose a secondary index architecture and user behavior enhanced vector space model(Behaviour Enhanced-VSM)for e-commerce search application scenarios,and then propose a BE-VSM-based text retrieval algorithm,and Applying this algorithm to Elastic Search,a real-time and efficient distributed e-commerce search engine is realized.Subsequently,this paper introduces the implementation scheme of the distributed e-commerce storage system,and performs functional and performance tests on the storage system.The test results show that the system interface is correct and complete,and the distributed file backup algorithm based on node QoS and interest can improve the file hit rate by 5%-10%compared with other algorithms.Compared with the IPFS official search method IPSE,the information search mechanism proposed in this paper can more than double the cumulative gain of user search results,which can provide highly reliable storage guarantee and real-time efficient search services for application systems.