计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (1): 94-101.DOI: 10.3778/j.issn.1002-8331.1701-0082

• 大数据与云计算 • 上一篇    下一篇

大数据下的CPS自适应路由服务

陈海华,张立臣   

  1. 广东工业大学 计算机学院,广州 510006
  • 出版日期:2018-01-01 发布日期:2018-01-15

Cyber-Physical System adaptive routing service with large data

CHEN Haihua, ZHANG Lichen   

  1. School of Computers, Guangdong University of Technology, Guangzhou 510006, China
  • Online:2018-01-01 Published:2018-01-15

摘要: 作为一个网络信息系统,信息物理融合系统(CPS)包含计算、人类活动以及通过消息通信技术实现的自动决策。CPS由许多异构的子系统组成,不同于其他控制系统,随着多种多样新型网络应用的涌现,传统的路由配置模式越来越难以适应CPS多样化的数据通信需求。因此,需要依据不同类型CPS子系统应用差异化的通信需求,在数据分组的传输路径上配置合适的路由功能,自适应地合成满足分组传输特性的路由服务,改善通信性能。根据大数据带来的数据间关联关系新范式,试图从大量的CPS应用通信流状态数据中,分析获取评判CPS通信服务质量与路由服务各属性之间的依赖关系,从而定制高效的路由服务。基于此,提出了大数据驱动的CPS自适应路由服务定制机制,主要由网内大量流状态数据为驱动,建立了CPS通信属性模型,挖掘CPS的通信性能对路由需求的依赖关系,获得候选路由功能集合。分析结果表明,提出的由大数据驱动的定制化路由服务可以有效地处理不断扩展的CPS网络通信需求。

关键词: 大数据, 信息物理融合系统(CPS), 定制化路由服务, 自适应路由服务

Abstract: As a network information system, Cyber-Physical System(CPS) includes computing, human activities and automatic decision which realizes mainly through message communication technology. Different from other control systems, CPS is composed of many heterogeneous subsystems. As long as the emergence of a variety of new network applications, the traditional routing configuration model becomes more and more difficult to adapt to the CPS’s diverse data communication needs. Therefore, it is necessary to provide appropriate communication demands on different types of CPS subsystems based on their specific communication, and compose routing services adaptively to satisfy their packet transmission characteristics with CPS’s fitness of communication service improved. Inspired by the new paradigm of data correlation brought by the big data, the dependencies between CPS’s communication adaptability and routing service properties are analyzed and exploited to help customize routing services efficiently, and a big date driven CPS’s adaptive routing service customization scheme is proposed. Mainly driven by the big data about the status of in-network flows, a CPS Communicate requirement attribute model is created, the dependencies between CPS’s communication performance and routing requirements are exploited, and then the candidate routing function set is obtained. The big data-driven, customized routing service proposed can effectively deal with the ever-expanding CPS network communication requirements.