Real-time performance optimization of TTE network based on "pay bursts only once" principle
-
摘要:
确定性通信的发展,促进了时间触发概念的引入。时间触发以太网(TTE)通过提供3种流量类别来支持混合安全性的实时应用:时间触发(TT)流量,具有完全的时间确定性;速率约束(RC)流量,具有确界的端到端延迟;尽力传(BE)流量。如何实现时间触发机制下RC流量实时性能的紧性分析,仍然是决定TTE网络顺利应用的开放式问题。在FIFO服务策略的假设下,将“一次性突发原则”的分析方法引入到TTE网络中,以观察该原则在时间触发网络性能分析中的影响。不同于航空电子全双工以太网(AFDX),具有更高优先级的TT流量会对RC流量的延迟分析产生关键影响,从而导致一次性突发分析的复杂性。通过建立聚合TT流量在及时阻断模式下的到达曲线模型,从而获得单条RC流量端到端的服务曲线模型,基于此实现了RC流量的最坏情况端到端延迟(WCD)评估,进一步完成了一次性突发原则下的分析对比。相较于已有工作,一次性突发原则可以得到RC流量更精确的最坏端到端延迟上界评估结果,有助于改善TTE网络性能评价紧性。通过A380拓扑组网案例的对比分析,相比于传统方法,所提方法RC流量平均延迟减少了12.05%。
-
关键词:
- 时间触发以太网(TTE) /
- 实时性分析 /
- 网络演算 /
- FIFO服务 /
- 延迟上界
Abstract:The development of deterministic communication promotes the introduction of time-triggered concepts. Time-Triggered Ethernet (TTE) network supports real-time applications of mixed safety by providing three traffic classes, including Time-Triggered (TT) traffic, which has complete time certainty, Rate-Constrained (RC) traffic, which has bounded end-to-end delay, and the Best-Effort (BE) traffic. How to realize the tight analysis of real-time performance of RC traffic under time-triggered mechanism is still an open problem that determines the smooth application of TTE network. Under the assumption of First Input First Output (FIFO) service, this paper introduces the analysis method of "pay bursts only once" principle into TTE network to observe the influence of this principle on the performance analysis of time-triggered network. Different from Avionics Full-Duplex Switched Ethernet (AFDX), TT traffic with higher priority will have a key impact on the delay analysis of RC traffic, resulting in the complexity of "pay bursts only once" principle analysis. In this paper, the arrival curve model of aggregated TT traffic in the timely blocking mode is established, and the end-to-end service curve model of single RC traffic is obtained. Based on this model, the Worst-Case end-to-end Delay (WCD) evaluation of RC traffic is realized. Compared with those in the existing studies, this method gives a tighter upper bound of the WCD for RC traffic, which is helpful to improve the tightness of TTE network performance evaluation. Through the comparative analysis of A380 topology networking cases, compared with the traditional method, the proposed method reduces the average delay of RC traffic by 12.05%.
-
Key words:
- Time-Triggered Ethernet (TTE) /
- real-time analysis /
- network calculus /
- FIFO service /
- delay upper bound
-
表 1 TC2用例实验结果对比
Table 1. Comparison of experimental results in TC2
TC2 TT流量/条 RC流量平均延迟/μs 平均延迟优化比例/% 本文方法 经典演算法 1 40 920.544 1 033.268 10.91 2 55 1 689.83 1 870.32 9.65 2 70 2 353.387 2 585.009 8.96 -
[1] Aerospace. Time-triggered Ethernet: SAE AS6802[S]. [S. l. ]: SAE International, 2011: 8-21. [2] KOPETZ H, ADEMAJ A, GRILLINGER P, et al. The time triggered Ethernet (TTE) design[C]//8th IEEE International Symposium on Object-Oriented Real-Time Distributed Computing. Piscataway: IEEE Press, 2005: 22-33. [3] IEEE. IEEE standard for Ethernet: IEEE 802.3[S]. [S. l. ]: LAN/MAN Standards Committee, 2012: 17-26. [4] LEE Y H, RACHLIN E, SCANDURA P A. Safety and certification approaches for Ethernet-based aviation databuses: DOT/FAA/AR-05/52[R]. [S. l. ]: Federal Aviation Administration, 2005. [5] ARINC. Aircraft data network, Part 7, Avionics full-duplex switched Ethernet network: ARINC 664P7[S]. [S. l. ]: Aeronautical Radio INC, 2009: 9-18. [6] STEINER W, BAUER G, HALL B, et al. TTEthernet dataflow concept[C]//8th IEEE International Symposium on Network Computing and Applications. Piscataway: IEEE Press, 2009: 319-322. [7] STEINBACH T, LIM H T, KORF F, et al. Tomorrow's in-car interconnect A competitive evaluation of IEEE 802.1 AVB and time-triggered Ethernet (AS6802)[C]//2012 IEEE Vehicular Technology Conference (VTC Fall). Piscataway: IEEE Press, 2012: 1-5. [8] SUEN J, KEGLEY R, PRESTON J. Affordable avionic networks with Gigabit Ethernet assessing the suitability of commercial components for airborne use[C]//2013 Proceedings of IEEE Southeastcon. Piscataway: IEEE Press, 2013: 1-6. [9] TǍMŞA S D, POP P, STEINER W. Design optimization of TTEthernet-based distributed real-time systems[J]. Real-Time Systems, 2015, 51(1): 1-35. doi: 10.1007/s11241-014-9214-8 [10] MAUCLAIR C, DURRIEU G. Analysis of real-time networks with Monte Carlo methods[J]. Progress in Flight Dynamics, Guidance, Navigation, Control, Fault Detection, and Avionics, 2013, 6: 501-514. [11] FRANCES F, FRABOUL C, GRIEU J. Using network calculus to optimize the AFDX network[J]. IEEE Transactions on Medical Imaging, 2006, 25(10): 1319-1328. doi: 10.1109/TMI.2006.880670 [12] BAUER H, SCHARBARG J L, FRABOUL C. Improving the worst-case delay analysis of an AFDX network using an optimized trajectory approach[J]. IEEE Transactions on Industrial Informatics, 2010, 6(4): 521-533. doi: 10.1109/TII.2010.2055877 [13] LI X T, SCHARBARG J, FRABOUL C. Improving end-to-end delay upper bounds on an AFDX network by integrating offsets in worst-case analysis[C]//IEEE Conference on Emerging Technologies and Factory Automation. Piscataway: IEEE Press, 2010: 1-8. [14] SCHARBARG J L, RIDOUARD F, FRABOUL C. A probabilistic analysis of end-to-end delays on an AFDX avionic network[J]. IEEE Transactions on Industrial Informatics, 2009, 5(1): 38-49. doi: 10.1109/TII.2009.2016085 [15] ADNAN M, SCHARBARG J L, ERMONT J, et al. Model for worst case delay analysis of an AFDX network using timed automata[C]//IEEE Conference on Emerging Technologies and Factory Automation. Piscataway: IEEE Press, 2010: 1-4. [16] STEINER W. Synthesis of static communication schedules for mixed-criticality systems[C]//14th IEEE International Symposium on Object/Component/Service-Oriented Real-time Distributed Computing Workshops. Piscataway: IEEE Press, 2011: 11-18. [17] TAMAS S D, POP P, STEINER W. Timing analysis of rate-constrained traffic for the TTEthernet communication protocol[C]//18th International Symposium on Real-Time Distributed Computing. Piscataway: IEEE Press, 2015: 119-126. [18] ZHAO L X, XIONG H G, ZHENG Z, et al. Improving worst case latency analysis for rate-constrained traffic in the time-triggered Ethernet network[J]. IEEE Communications Letters, 2014, 18(11): 1927-1930. doi: 10.1109/LCOMM.2014.2358233 [19] ZHAO L, POP P, LI Q J, et al. Timing analysis of rate-constrained traffic in TTEthernet using network calculus[J]. Real-Time Systems, 2017, 53(2): 254-287. doi: 10.1007/s11241-016-9265-0 [20] 何锋. 机载网络技术基础[M]. 北京: 国防工业出版社, 2018.HE F. Fundamentals of airborne network[M]. Beijing: National Defense Industry Press, 2018(in Chinese). [21] STEINER W, DUTERTRE B. SMT-based formal verification of a TTEthernet synchronization function: In formal methods for industrial critical systems[M]. Berlin: Springer, 2010. [22] FIDLER M. Extending the network calculus pay bursts only once principle to aggregate scheduling[C]//2nd International Workshop on Quality of Service in Multiservice IP Networks. Berlin: Springer, 2003: 19-34. [23] LENZINI L, MARTORINI L, MINGOZZI E, et al. Tight end-to-end per-flow delay bounds in FIFO multiplexing sink-tree networks[J]. Performance Evaluation, 2006, 63(9-10): 956-987. doi: 10.1016/j.peva.2005.10.003 [24] BOYER M, FRABOUL C. Tightening end to end delay upper bound for AFDX network calculus with rate latency FIFO servers using network calculus[C]//IEEE International Workshop on Factory Communication Systems. Piscataway: IEEE Press, 2008: 11-20. [25] CRUZ R L. A calculus for network delay, Part Ⅰ, Network elements in isolation[J]. IEEE Transactions on Information Theory, 1991, 37(1): 114-131. doi: 10.1109/18.61109 [26] LE B J, THIRAN P. Network calculus: A theory of deterministic queuing systems for the internet[M]. Berlin: Springer, 2001.