ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展 ›› 2019, Vol. 56 ›› Issue (3): 453-466.doi: 10.7544/issn1000-1239.2019.20170667

• 网络技术 •    下一篇

EasiDARM:基于分布式的物联网设备自适应注册方法

施亚虎1,2,石海龙1,崔莉1   

  1. 1(中国科学院计算技术研究所 北京 100190); 2(中国科学院大学 北京 100049) (shiyahu@ict.ac.cn)
  • 出版日期: 2019-03-01
  • 基金资助: 
    国家自然科学基金项目(61502461,61672498)

EasiDARM: Distributed Based Adaptive Register Method for Internet of Things

Shi Yahu1,2, Shi Hailong1, Cui Li1   

  1. 1(Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190); 2(University Chinese Academy of Sciences, Beijing 100049)
  • Online: 2019-03-01

摘要: 随着物联网水平化接入协议的逐步成熟与实用化,将设备接入云平台以对设备进行实时访问逐步成为一种主流架构.由于物联网设备或其所处网络环境存在动态性,为了保证物联网云平台与设备之间的双向实时访问,设备往往需要到云平台进行周期注册.而现有注册方法存在开销大、耗时长的缺点,当大量设备或网络处于强动态场景时,云平台的注册开销将急剧加大.为此,提出了基于分布式的物联网设备自适应注册方法(EasiDARM),将复杂且耗时的周期探测过程分配给多个设备执行,并通过参数同步实现结果实时共享,极大地加快周期探测和设备自适应过程,降低物端、网端和云端开销.该方法将周期探测过程分为“快更新”和“快收敛”前后2个阶段,“快更新”阶段采用指数增长方式进行任务分配及周期探测,加快注册周期增长速度,“快收敛”阶段采用线性增长方式进行任务分配及周期探测,加快周期探测收敛速度.实验结果表明:较于传统的自适应注册方法,该方法周期探测耗时能减少46%,物端开销能降低46%,网端通信开销和云端处理开销能降低53%;同时能使云平台的突发性开销降低36%.

关键词: 物联网, 海量设备, 设备自适应注册, 网络地址转换, 任务分配

Abstract: With the perfecting and practicability of the access protocols of the Internet of things (IoT), it is becoming a mainstream architecture to connect devices to cloud platforms for support of real-time access. Considering the dynamic nature of IoT devices and networks, we must make the devices register on the platforms periodically to ensure real-time access between the devices and the platforms. However, existing register methods cost much time and consume a lot of resources on devices, Internet and platforms; and what’s more, the platforms will be overload when many devices or networks are dynamic. In this paper, we present a distributed based adaptive register method (EasiDARM) to reduce the costs of resources and time. By using multiple devices to complete the measurement of network address translation (NAT) timeout which is complex and time-consuming, and sharing the result of measurement among the devices, EasiDARM accelerates the processes of the measurement and self-adaption. The process of measurement is divided into fast update (FU) and fast converge (FC) stage, EasiDARM assigns and probes the candidate intervals with exponential growth during the FU stage to extend the register interval rapidly, and with linear growth during the FC stage to converge the measurement quickly. Through experiments show that compared with the traditional adaptive register method, EasiDARM can reduce 46% of time, 46% of consumption on devices and 53% of consumption on Internet and platforms in the process of measurement, and cut 36% of instantaneous consumption on platforms.

Key words: Internet of things, massive device, device adaptive register, network address translation (NAT), task distribution

中图分类号: