摘要
提出多路数据流计算模型和对云计算物理系统的分层模块化建模方法.云计算物理系统包含大量相连的嵌入式设备和计算节点.其开发难点在于系统复杂性所带来的设计和管理困难,以及对海量数据的处理.可行的解决办法是以多路数据流构造抽象的设计模型,利用自动化工具进行仿真、检验和优化,再通过映射生成最终系统.多路数据流扩展了传统数据流和Google的MapReduce计算框架,在简化设计的同时也使最终系统易于在云计算环境中实现.
The multi-way dataflow model of computation was invented, and an approach to hierarchical compositional modeling was proposed for cloud cyber-physical systems. A cloud cyber-physical system comprises a large quantity of interconnected embedded devices and computer nodes. Development difficulty lies in the great complexity causing challenges in design and management, and in the massive data size. A feasible solution is to construct abstract design models with multi-way dataflow, to utilize automation tools for simulation, verification and optimization, and to map designs to final systems. Multi-way dataflow extends traditional dataflow and the MapReduce computation framework of Google. It simplifies design, and also facilitates the implementation of final systems in cloud computing environments.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2013年第6期1570-1576,共7页
Systems Engineering-Theory & Practice
