隨著自動駕駛和物聯網等技術的迅速發展，分散式即時通訊系統對動態負載適應性和跨平台一致性提出更高的要求。現有基於Data Distribution Service (DDS) 的分散式系統通常使用靜態配置，無法根據實際負載動態調整發布頻率，且在異構平台部署時容易出現顯著的性能差異，尤其是在高負載條件下容易出現嚴重性能退化。
本研究實現了一個基於FastDDS的跨平台分散式自適應頻率控制系統，採用雙層通訊架構：DDS負責高頻數據傳輸，HTTP RESTful API負責控制指令傳遞，實現動態傳輸速率調整機制。藉由在Subscriber端即時監測資料接收率，計算後動態調整Publisher的傳輸速率。系統包含三個核心組件：SubscriberReporter負責收集即時監測資料接收率與上報，透過HTTP POST回報至中央控制器；RateController通過計算得到新的傳輸頻率；PublisherAdjuster執行頻率調整指令，進行實時動態調整。該系統有效地提高了系統整體效能與穩定性。
實驗在真實異構環境中進行，包含Linux發布者節點、Ubuntu控制節點和Windows訂閱者節點。設計了全因子實驗，涵蓋3個負載等級 (3000/5000/7000 msg/s)、3種QoS配置組合 (RELIABLE-RELIABLE、RELIABLE-BEST_EFFORT、BEST_EFFORT-BEST_EFFORT) 和2個接收平台，共18個實驗場景。
實驗結果表明，自適應頻率控制器在所有18個場景中都實現了正向改善，成功率達100%。整體平均改善幅度為13.22%，最大改善效果達到69.90%。在極端場景下，控制器展現出強大的系統恢復能力，能將準確率從0.40%的幾乎完全失效狀態恢復到56.58%。跨平台性能分析發現，Ubuntu平台表現出較好的負載承受能力和穩定性，而Windows平台在RELIABLE QoS模式下容易出現性能不穩定，但控制器能有效緩解這種差異。

The rapid development of technologies such as the Internet of Things (IoT), and autonomous driving has led to higher requirements for decentralized real-time communication systems in terms of dynamic load adaptability and cross-platform consistency. Current decentralized systems based on Data Distribution Service (DDS) typically use static configurations that cannot dynamically adjust the dissemination frequency according to the actual load. These systems are also prone to significant performance differences when deployed on heterogeneous platforms, especially under high load conditions.
This study presents a cross-platform, decentralized, adaptive frequency control system based on FastDDS. It adopts a two-layer communication architecture: DDS is responsible for high-frequency data transmission and the HTTP RESTful API controls instruction delivery to adjust the transmission rate dynamically. Real-time monitoring of the data reception rate on the subscriber side and subsequent calculation dynamically adjust the publisher's transmission rate. The system consists of three core components: The SubscriberReporter collects and reports performance data to the central controller via HTTP POST, the RateController calculates the new publishing frequency, and the PublisherAdjuster executes frequency adjustment commands for real-time dynamic adjustment. This system improves the overall performance and stability.
Experiments were conducted in a real heterogeneous environment that included a Linux publisher node, an Ubuntu control node, and a Windows subscriber node. The fully factorial experiments covered three load levels (3,000, 5,000, and 7,000 messages per second), three QoS configuration combinations (RELIABLE-RELIABLE, RELIABLE-BEST_EFFORT, and BEST_EFFORT-BEST_EFFORT), and two testbeds, for a total of 18 experimental scenarios.
The experimental results show that the adaptive frequency controller improves performance in all 18 scenarios, achieving a 100% success rate. The average overall improvement is 13.22%, with a maximum improvement of 69.90%. In extreme scenarios, the controller demonstrated robust system recovery, improving system reliability from near-total failure (0.40%) to 56.58% and achieving 141x performance recovery. Cross-platform performance analysis reveals that the Ubuntu platform exhibits greater load tolerance and stability. In contrast, the Windows platform is susceptible to performance vulnerability in RELIABLE QoS mode; however, the controller effectively mitigates this difference.

References
[1] G. Pardo-Castellote, "Omg data-distribution service: Architectural overview," in 23rd International Conference on Distributed Computing Systems Workshops, 2003. Proceedings., 2003, pp. 200-206: IEEE.
[2] Real Time Innovations. (2025). Data Distribution Service (DDS) Community-RTI Connext Users. Available: https://community.rti.com/documentation
[3] eProsima. (2025). eProsima Fast DDS. Available: https://fast-dds.docs.eprosima.com/en/latest/
[4] OpenDDS Foundation. (2025). OpenDDS. Available: https://opendds.readthedocs.io/en/latest-release/
[5] Eclipse Foundation. (2025). Eclipse Cyclone DDS. Available: https://cyclonedds.io/docs/
[6] ADLINK Technology Inc. (2025). Vortex OpenSplice DDS. Available: https://www.adlinktech.com/en/dds-community-software-evaluation
[7] Twin Oaks Computing, Inc. (2025). CoreDX DDS. Available: https://www.twinoakscomputing.com/coredx/documentation
[8] S. Jeong, T. Ga, I. Jeong, and J. Choi, "Behavior tree driven multi-mobile robots via data distribution service (DDS)," in 2021 21st International Conference on Control, Automation and Systems (ICCAS), 2021, pp. 1633-1638: IEEE.
[9] M. L. Gambo, A. Danasabe, B. Almadani, F. Aliyu, A. Aliyu, and E. J. R. Al-Nahari, "A Systematic Literature Review of DDS Middleware in Robotic Systems," Robotics, vol. 14, no. 5, p. 63, 2025.
[10] Y. Park, D. Lim, D. Min, S. A. Kim, "Research on Design of DDS-based Conventional Railway Signal Data Specification for Real-time Railway Safety Monitoring and Control," Journal of the Korea Institute of Information and Communication Engineering, vol. 20, no. 4, pp. 739-746, 2016.
[11] J. So, K.-H. Shin, and J. Ahn, "A Study on DDS (Data Distribution Service) Application for Real-time Monitoring and Control in Operation Console of the Railway Safety Control Platform," Journal of The Korean Society For Urban Railway, vol. 6, no. 4, pp. 279-286, 2018.
[12] M. Essers and T. H. Vaneker, "Evaluating a prototype approach to validating a DDS-based system architecture for automated manufacturing environments" Procedia CIRP, vol. 25, pp. 385-392, 2014.
[13] B. Almadani and S. M. Mostafa, "IIoT based multimodal communication model for agriculture and agro-industries," IEEE Access, vol. 9, pp. 10070-10088, 2021.
[14] G. Yoon, S. Lee, and H. Choi, "Qos optimizer," in 2016 International conference on platform technology and service (PlatCon), 2016, pp. 1-5: IEEE.
[15] R. S. Auliya, C.-C. Chen, P.-R. Lin, D. Liang, and W. J. Wang, "Optimization of message delivery reliability and throughput in a DDS-based system with per-publisher sending rate adjustment," Telecommunication Systems, vol. 84, no. 2, pp. 235-250, 2023.
[16] Z. Meng et al., "Enabling high quality {Real-Time} communications with adaptive {Frame-Rate}," in 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23), 2023, pp. 1429-1450.
[17] B. Almadani and S. Abudalfa, "QoS adaptation for publish/subscribe middleware in real-time dynamic environments," International Arab Journal of Information Technology (IAJIT), vol. 14, no. 2, 2017.
[18] A. Baunthiyal, M. G. Rakesh, A. K. Jain, and E. Technology, "Criteria Set for Evaluation of different DDS Distributions," International Journal for Research in Applied Science and Engineering Technology, vol. 9, no. 1, pp. 119-128, 2021.
[19] Y. Fu, L. Hao, and D. Guo, "Application research of distributed simulation system based on data distribution," in 2019 IEEE international conference on unmanned systems and artificial intelligence (ICUSAI), 2019, pp. 268-273: IEEE.