高并发数据共享系统基于Push-Pull混合模式的性能优化  

The Performance Optimization Based on Push-Pull Mixed Mode in the High Concurrent Data Sharing System

作　　者：王子光[1] 王子明 WANG Ziguang;WANG Ziming(The People’s Insurance Company（Group）of China Limited,Beijing 100084,China;CITIC Group,Beijing 100004,China)

机构地区：[1]中国人民保险集团股份有限公司,北京100084 [2]中国中信集团有限公司,北京100004

出　　处：《现代信息科技》2018年第4期65-67,共3页Modern Information Technology

摘　　要：随着面向服务架构的大规模分布式系统的应用,多个系统之间由于数据集成的需要,彼此间要高并发传递大量数据,特别是状态数据的同步,对数据的实时性要求越来越高。数据单纯通过客户端以Pull(拉取)模式获取已经不能满足实时性要求,高频率服务调用也会给服务端数据库带来较大的压力。单纯通过服务端以Push(推送)模式推送给客户端也不能满足客户端对个性化数据的需求,大量推送既会给服务端带来较大压力,又会造成客户端数据处理不及时。本文针对高并发数据共享系统应用过程中产生的性能问题,提出一种基于pub/sub消息处理的Push-Pull混合模式优化方案,实践表明该方案极大地提高了系统可用性及数据安全性,对同类系统的性能优化具有较好的借鉴作用。With the application of large-scale distributed systems based on service oriented architecture,due to the need of data integration between multiple systems,these systems transfer large amounts of data between each other,especially the synchronous state of the data. The data obtained by the client through pull mode cannot meet the real-time requirements,high-frequent service-calls have also bring greater pressure to the database server through push mode;only by the server through the push mode to the client,it cannot meet the needs of personalized data,also a large number of push will bring greater pressure on both server and client which cannot cause data processing timely. In this paper,with the performance problems in the application of the high-concurrent data sharing system,it put forward a kind of Push-Pull mixed mode optimization scheme based on pub/sub message processing,the practice shows that the scheme greatly improves the system availability and data security,also provides a good reference for optimizing the performance of similar systems.

关 键 词：高并发 数据共享系统 性能优化 pub/sub消息处理 Push-Pull混合模式 

分 类 号：TN929.5[电子电信—通信与信息系统]