航空学报 > 2022, Vol. 43 Issue (7): 325578-325578   doi: 10.7527/S1000-6893.2021.25578

优先级驱动的泛化航电网络实时性能分析

于思凡, 何锋, 熊华钢   

  1. 北京航空航天大学 电子信息工程学院, 北京 100083
  • 收稿日期:2021-03-29 修回日期:2021-09-07 发布日期:2021-05-26
  • 通讯作者: 何锋,E-mail:robinleo@buaa.edu.cn E-mail:robinleo@buaa.edu.cn
  • 基金资助:
    国家自然科学基金(62071023)

Priority-driven generalized real-time performance analysis of avionics network

YU Sifan, HE Feng, XIONG Huagang   

  1. School of Electronics and Information Engineering, Beihang University, Beijing 100083, China
  • Received:2021-03-29 Revised:2021-09-07 Published:2021-05-26
  • Supported by:
    (]National Natural Science Foundation of China (62071023)

摘要: 交换式网络利用空分交换结构大大提高了网络的整体通信容量,已成为新一代航空电子系统组网互联的典型手段,可以采用泛化网络演算实施典型组网特性下的机载网络实时性能画像和预测,但缺乏优先级策略下的网络实时性能分析手段。在对机载交换式网络优先级控制策略特征参数抽象化的基础上,利用网络演算理论实施了高低优先级控制的泛化网络演算建模,在突发度包络函数的边界约束下完成了不同优先级流量占比的端到端确界延迟分析,并给出了优先级占比分配方案推荐。在典型机载网络拓扑结构下利用随机生成案例对优先级驱动的泛化网络演算模型进行实验验证。结果表明:相比于采用详细流量配置的演算结果,本文提出的泛化网络演算的有效性超过33%,实现了基于组网特征的机载网络性能预估。

关键词: 交换式网络, 优先级, 组网特征, 网络演算, 延迟分析, 实时性

Abstract: Switching networks make use of the structure of air separation switching to increase the overall communication capacity remarkably, and have therefore become a typical means for interconnecting new generation avionics systems. Generalized network calculus can be used to predict the real-time performance of switching networks according to typical network features. However, generalized network calculus lacks the means to analyze the real-time performance of the networks which distinguish high and low priority. On the basis of abstracting the features of the priority control strategy for switched networks, the priority-driven generalized network calculus models the high and low priority streams using the network calculus theory. Under the boundary constraint of the burst envelope function, the end-to-end delay-bound analysis of different priority proportions is completed, and the priority ratio scheme is given. In the typical airborne network topology, the model proposed is verified experimentally with a randomly generated case. The results show that compared with the calculus using detailed traffic configuration, the calculus proposed is 33% more effective, and can realize performance prediction of airborne switched networks based on network features.

Key words: switching network, priority, network feature, network calculus, delay analysis, real-time perfomancehttp

中图分类号: