图计算框架的大规模社交网络数据处理  

Large-scale social network data processing using graph computing framework

作　　者：党莉莉 于强 张晔 DANG Lili;YU Qiang;ZHANG Ye(Henan Branch of China Telecom,Zhengzhou 450000,China)

机构地区：[1]中国电信股份有限公司河南分公司,郑州450000

出　　处：《计算机应用文摘》2025年第3期157-159,共3页

摘　　要：随着社交网络规模的指数级增长,传统数据处理方法面临挑战。基于图计算框架在处理大规模社交网络数据中的应用情况,文章分析了Pregel,GraphX和PowerGraph等主流框架的优缺点。研究表明,基于BSP模型的分布式图计算框架能显著提升处理效率,而针对社交网络特点的改进图分区算法能够有效降低跨分区通信开销。With the exponential growth in the size of social networks,traditional data processing methods face challenges.Focusing on the application of graph computing frameworks in processing large-scale social network data,this paper analyzes the advantages and disadvantages of mainstream frameworks such as Pregel,GraphX and PowerGraph.The research shows that the distributed graph computing framework based on BSP model can significantly improve the processing efficiency,and the improved graph partitioning algorithm based on the characteristics of social networks can effectively reduce the cross-partition communication overhead.

关 键 词：图计算框架 社交网络分析 大规模数据处理 分布式计算 图分区 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]