一种通用混合计算中间件的设计与实现  

Design and implementation of a general-purpose hybrid computing middleware

作　　者：张洪豪[1,2] 姚欣[1,2] 王劲松 姚世春[3] ZHANG Honghao;YAO Xin;WANG Jinsong;YAO Shichun(School of Computer and Communication Engineering,Tianjin University of Technology,Tianjin 300384,China;Tianjin Key Laboratory of Intelligent Computing&Novel Software Technology,Tianjin University of Technology,Tianjin 300384,China;ANTC Technology(Tianjin)Co.,Ltd.,Tianjin 300230,China)

机构地区：[1]天津理工大学计算机与通信工程学院,天津300384 [2]天津理工大学天津市智能计算及软件新技术重点实验室,天津300384 [3]安态诚科技(天津)有限公司,天津300230

出　　处：《现代电子技术》2020年第12期55-60,共6页Modern Electronics Technique

基　　金：国家重点研发计划(2018YFC0831405);国家自然科学基金面上项目(61673295);天津市自然科学基金项目(18JCYBJC85200)天津市自然科学基金项目(18JCZDJC30700)。

摘　　要：随着接入设备的激增,网络应用及协议的不断涌现和互联网流量的爆发式增长,给数据分析与处理带来了极大的挑战。大数据环境中数据处理应用所呈现出的趋势是实时在线计算、离线批式计算以及动态响应式计算相互结合、相互影响,这就要求底层计算资源具备同时提供实时计算、批式计算及响应式计算的能力,然而传统的系统架构和编程模型已经无法满足这种混合计算需求。文中设计并实现一种构建在流式数据处理过程IStream之上的混合计算中间件(Generalpurpose Hybrid Computing Middleware,GHCM),它以统一、灵活的高层抽象将相互依赖的高并发实时处理业务逻辑、批处理业务逻辑和跨层次动态调用无缝融合起来。实验数据表明,GHCM在解决混合计算问题上具有良好的性能,在大幅缩短开发时间的同时为用户提供了可重用和易扩展等支持。The sharp increase of access devices,the continuous emergence of network applications and protocols,and the explosive growth of Internet traffic bring a great challenge to data analysis and processing. The trend presented by the data processing applications in the big data environment is the mutual integration and mutual interaction of real-time online computing,offline batch computing and dynamic responsive computing,which requires that the underlying computing resources have the ability to simultaneously provide real-time computing,batch computing and responsive computing,but the traditional system architecture and programming model is unable to meet the needs of this kind of hybrid computing. A general-purpose hybrid computing middleware(GHCM) based on the streaming data treating process IStream is designed and implemented,which integrates the interdependent high concurrent real-time processing business logic,batching business logic and cross level dynamic call seamlessly with a unified and flexible high-level abstraction. The experimental data show that GHCM has good performance in solving the hybrid computing difficulty,and provides users with reusable and easily extensible support while significantly reducing development time.

关 键 词：混合计算中间件 IStream模型设计 数据处理 程序无缝融合 系统实现 性能分析 

分 类 号：TN911.2-34[电子电信—通信与信息系统] TP302[电子电信—信息与通信工程]