分布式实时系统数据分发服务DDS技术研究

doi:10.16180/j.cnki.issn1007-7820.2020.08.007

摘要/Abstract

摘要：

数据分发服务DDS是面向分布式实时系统的中间件技术规范。随着分布式实时系统复杂度和规模的增大,系统对数据传输的质量要求也在大幅提高。DDS采用以数据为中心的发布/订阅模式,能够有效满足各种分布式应用的性能要求和硬实时要求。文中围绕数据分发服务技术展开讨论,分析对比了4种常见的数据分发模型,指出了各自的优缺点。文中围绕通信机制、服务质量、传输框架、发现过程以及标准实现等5个方面对DDS规范进行了阐述;从DDS配置的复杂性、广域网应用DDS以及无线网络应用DDS共3个方面重点介绍了DDS应用面临的挑战,并探讨了未来的研究方向。

关键词: 数据分发服务, 中间件, 数据分发模型, 以数据为中心的发布/订阅, 分布式实时系统, 服务质量

Abstract:

Data Distribution Service is the technical specification of communication middleware for distributed real-time system. With the increase of complexity and scale of distributed real-time system, the quality requirements of data transmission in the system are greatly increased. DDS adopts data-centric publish/subscribe mode, which meet the performance requirements and hard real-time requirements of various distributed applications. The article firstly analyzed and compared four common data distribution models and pointed out their respective advantages and disadvantages. Then, the DDS specification was elaborated in five aspects: communication mechanism, quality of service, transmission framework, discovery process and standard implementation. The challenges faced by DDS applications were reviewed from three aspects: the complexity of DDS configuration, DDS application in WAN and DDS application in wireless network, and the future research directions are discussed.

Key words: data distribution service, middleware, data distribution model, data-centric publish-subscribe, distributed real-time system, quality of service

中图分类号:

TP393

王天一,高博. 分布式实时系统数据分发服务DDS技术研究[J]. 电子科技, 2020, 33(8): 40-45.

WANG Tianyi,GAO Bo. Research on Data Distribution Service of Distributed Real-time System[J]. Electronic Science and Technology, 2020, 33(8): 40-45.

图/表 4

参考文献 21

[1]	任昊利, 李旺龙, 张少扬, 等. 数据分发服务——以数据为中心的发布/订阅式通信[M]. 北京: 清华大学出版社, 2014.
	Ren Haoli, Li Wanglong, Zhang Shaoyang, et al. Data distribution service——data-centric publish/subscribe communication[M]. Beijing: Tsinghua University Press, 2014.
[2]	OMG. Data distribution service for real-time systems specification version1.4[EB/OL]. (2015-03-02) [2018-11-12] http://www.omg.org/spec/DDS/1.4.
[3]	朱华勇, 张庆杰, 沈林成, 等. 分布式系统实时发布/订阅数据分发服务[M]. 北京: 国防工业出版社, 2013.
	Zhu Huayong, Zhang Qingjie, Shen Lincheng, et al. Distributed system real-time publish/subscribe data distribution service[M]. Beijing: National Defense Industry Press, 2013.
[4]	OMG. Deployment and configuration of component based distributed applications version4.0[EB/OL]. (2006-04-03) [2018-11-12], http://www.omg.org/spec/DEPL/4.0.
[5]	Sanchez-Monedero J, Povedano-Molina J, Lopez-Vega J M, et al. Bloom filter-based discovery protocol for DDS middleware[J]. Journal of Parallel & Distributed Computing, 2011,71(10):1305-1317.
[6]	Dianes J A, Díaz M, Rubio B. Using standards to integrate soft real-time components into dynamic distributed architectures[J]. Computer Standards & Interfaces, 2012,34(2):238-262.
[7]	An K, Kuroda T, Gokhale A S, et al. Model-driven generative framework for automated OMG DDS performance testing in the cloud[C]. Indianapolis: Companion Publication for Conference on Systems,ACM, 2013.
[8]	Feiock D, Hill J H. Using component-based middleware to design and implement Data Distribution Service (DDS) systems[C]. Santander:Software Engineering & Advanced Applications,IEEE, 2013.
[9]	Tekinerdogan B, Turgay Çelik, Ömer Köksal. Generation of feasible deployment configuration alternatives for data distribution service based systems[J]. Computer Standards & Interfaces, 2018(4):59-68.
[10]	Lopez Vega J M, Povedano Molina J, Pardo Castellote G, et al. A content-aware bridging service for publish/subscribe environments[J]. Journal of Systems and Software, 2013,86(1):108-124. doi: 10.1016/j.jss.2012.07.033
[11]	Lee K, Kim C, Kim K T, et al. Router design for DDS: Architecture and performance evaluation[C]. Bangkok:International Conference on Big Data & Smart Computing,IEEE Computer Society, 2014.
[12]	Hakiri A, Berthou P, Gokhale A, et al. Supporting end-to-end quality of service properties in OMG data distribution service publish/subscribe middleware over wide area networks[J]. Journal of Systems and Software, 2013,86(10):2574-2593. doi: 10.1016/j.jss.2013.04.074
[13]	Hakiri A, Berthou P, Gokhale A, et al. Supporting SIP-based end-to-end data distribution service QoS in WANs[J]. Journal of Systems and Software, 2014,95(6):100-121. doi: 10.1016/j.jss.2014.03.078
[14]	Bertaux L, Hakiri A, Medjiah S, et al. A DDS/SDN based communication system for efficient support of dynamic distributed real-time applications[C]. Toulouse:IEEE/ACM 18^th International Symposium on Distributed Simulation and Real Time Applications,IEEE , 2014.
[15]	Beckmann K, Thoss M. A wireless sensor network protocol for the OMG Data Distribution Service[C]. Klagenfurt: Intelligent Solutions in Embedded Systems,IEEE, 2012.
[16]	Park H S, Jang J H, Kim J D, et al. Bandwidth-aware DDS communication in WLAN[C]. Yantai:International Conference on Systems & Informatics,IEEE, 2012.
[17]	Boonma P, Suzuki J. Self-configuring publish/subscribe middleware for wireless sensor networks[C]. Las Vegas: Consumer Communications & Networking Conference, 2009.
[18]	Boonma P, Suzuki J. Toward interoperable publish/subscribe communication between wireless sensor networks and access networks[C]. Las Vegas:Consumer Communications and Networking Conference, 2009.
[19]	Beckmann K, Dedi O. DsDDS:A portable data distribution service implementation for WSN and IoT platforms[C]. Ancona:International Workshop on Intelligent Solutions in Embedded Systems, 2015.
[20]	周莹. 受限带宽环境下实时数据分发服务传输优化技术研究[D]. 北京:中国舰船研究院, 2016.
	Zhou Ying. Research on data distribution service transmission optimization in limited bandwidth networks environments[D]. Beijing:China Ship Research Institute, 2016.
[21]	Kang W, Kim J. PDDS:scalable sensor data distribution for cyber-physical systems using GPGPUs[J]. IEEE Internet of Things Journal, 2018(7):133-141.

供应商及产品	类型	传输模型	结构	特性
OCI公司 Open DDS	开源	默认:单播多播	有中心体系结构	实现和配置简单;节点失效、存在性能瓶颈。
RTI公司 Connext DDS	商用	默认:单播多播	分散体系结构	时延和抖动较小;节点失效。
PrismTech 公司 OpenSplice DDS	商用	默认:广播多播	联邦制体系结构	不存在节点失效;配置复杂、移植性较差。