北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2023, Vol. 46 ›› Issue (3): 37-42.

• 人工智能使能网络通信 • 上一篇    下一篇

确定性网络跨域传输架构与DRL流量调度算法

谭炜骞1,2,吴斌伟2,汪硕2,3   

  1. 1. 东南大学 网络空间安全学院 2. 网络通信与安全紫金山实验室 3. 北京邮电大学 网络与交换技术国家重点实验室
  • 收稿日期:2022-07-28 修回日期:2022-11-28 出版日期:2023-06-28 发布日期:2023-06-05
  • 通讯作者: 吴斌伟 E-mail:wubinwei@ pmlabs.com.cn
  • 基金资助:

    国家重点研发计划项目(2020YFB1806406)

Cross-Domain Deterministic Networking Architecture and DRL Flow Scheduling

TAN Weiqian1,2, WU Binwei2, WANG Shuo2,3   


  • Received:2022-07-28 Revised:2022-11-28 Online:2023-06-28 Published:2023-06-05

摘要:

针对时间敏感网络的跨域传输问题,提出了确定性跨域传输架构和基于深度强化学习(DRL)的调度算法确定性跨域传输是一种融合循环队列转发与确定性网际互连协议的广域确定性组网架构,通过定义跨域周期映射函数,建立基于时隙的确定性传输通道,保障有界的传输时延;基于 DRL 的时隙路径联合在线调度算法,定义 DRL状态动作和奖励,以收益和最大化为目标对不同收益的业务流进行调度实验结果表明,所提跨域传输架构和调度算法可以保障端到端的确定性传输,显著提高了流量调度收益值,以保障重要流量的传输

关键词:

Abstract:

A deterministic cross-domain transmission architecture and a deep reinforcement learning(DRL)-based scheduling algorithm are proposed to address the cross-domain transmission problem in time-sensitive networks. The deterministic cross-domain transmission architecture is a wide-area deterministic networking architecture that integrates cyclic queuing forwarding with deterministic Internet Protocol. By defining a cross-domain cycle mapping function, a time-slot-based deterministic transmission channel is established to ensure bounded transmission delay. DRL states, actions, and rewards are detined in the DRL-based time-slot path joint online scheduling algorithm, and the scheduling target is to maximize the total earning value of all scheduled flows with different earning values. Experimental results demonstrate that the proposed cross-domain transmission architecture and scheduling algorithm can ensure end-to-end deterministic transmission, significantly improve the earning value of traffic scheduling, and guarantee the transmission of important traffic.

Key words: deterministic networking, large-scale cross-domain communication, flow scheduling, deep reinforcement learning

中图分类号: