Design And Implementation Of Graph Computing Platform Framework Based On Graph Database

With the era of full-scale mobile Internet and big data now,fields such as social networks and knowledge graphs have received more and more research and attention.The data in these fields have a strong correlation with each other.When the traditional relationship database is processing related data,a large number of table connections are required,which causes inefficiency and a large time cost.Faced with the problems caused by such application scenarios,graph database technology has been produced and developed rapidly.Neo4j,as the top graph database in the market and Spark GraphX in the field of graph computing,has been more and more widely used in this context.However,many graph databases also bring inconvenience to developers,so this article hopes to design a unified and extensible graph computing platform framework that can cover different graph databases at the bottom,and unify the upper application and display interface.This article takes graph database Neo4j and graph calculation engine Spark GraphX as examples to design and implement a graph computing platform framework.The framework is divided into persistent modules,algorithm modules,control modules and application modules at the overall design level.At the module implementation code level,simple factories,factory methods,and abstract factory design patterns are introduced to allow the entire framework to understand the coupling and scalability.This paper implements the interface of the framework's persistent module and some algorithm modules,including graph data generator,Neo4j importer and HDFS importer,as well as graph algorithms such as SSSP,PageRank,node centrality,triangle count,etc.,and through D3.js The framework visualizes the calculation results.Finally,the graph computing platform framework designed and implemented in this paper is applied to the recommendation system,which implements the bipartite graph recommendation based on the PersonRank algorithm,and further applies the graph database Neo4j and graph computing engine Spark GraphX technology to the actual scene.