通訊網路自適應之加強比例積分控制器主動式佇列管理演算法

A-PIE: An Adaptive Active Queue Management Algorithm for Communication Networks

林信志(Hsin-Chih Lin)

指導教授 ： 高榮鴻

國立交通大學/電機學院/電機學院電信學程/碩士(2018年)

摘要

主動式佇列管理（AQM）在現今網路複雜的應用情況下，相較於傳統的佇列處理方式，能提供使用者更佳的使用經驗與傳輸品質。而加強比例積分控制器（ PIE）這個近年提出的 AQM 演算法，相較於其他 AQM 演算法有許多的優點，在模擬中也表現出相當優良的封包延遲管理能力與通道利用率。但是，雖然有許多參數已經具備自動調節的功能， PIE 的主要參數卻需要在使用之前預先設定並且無法在運作中修改，這造成了 PIE 在實際應用上非常大的限制。在這篇論文中，我們提出 PIE 的改進策略，強調 PIE 的佇列延遲設定需要參考網路環境做調整，並利用簡化後的封包來回時間（ RTT）偵測方法，來協助達成調整的目的。最後，我們利用網路模擬工具（ ns-2）模擬改進後的 PIE 演算法，並分析這個演算法所達成的效能改善。

關鍵字

主動式佇列管理演算法 ； 加強比例積分控制器

並列摘要

Active queue management (AQM) could mitigate the bufferbloat problem and enhance the quality of experience for network applications. In comparison with other AQM algorithms, the Proportional Integral Controller Enhanced (PIE) algorithm has been shown to better manage the queuing delay without sacrificing channel efficiency. However, when deploying the PIE algorithm to a network equipment, key parameters such as Reference Queuing Delay have to be tuned before operation and cannot be modified afterwards. In this thesis, we address the flaws of the PIE algorithm and propose a novel adaptive PIE algorithm. We observe that the Reference Queuing Delay should adapt to the network state. In addition, we propose the partial Round Trip Time (RTT) estimation method for queue length adjustment. Furthermore, we use the well-known ns-2 network simulator to evaluate the revised PIE algorithm. Our study shows that the proposed A-PIE algorithm significantly outperforms the original PIE algorithm.

並列關鍵字

AQM ； PIE ； Bufferbloat

參考文獻

[1] G. White and D. Rice, “Active queue management in DOCSIS 3.x cable modems,” CableLabs Technical Report, 2014.

[2] M. Dischinger, A. Haeberlen, K. P. Gummadi, and S. Saroiu, “Characterizing residential broadband networks,” in Internet Measurement Comference, pp. 43–56, 2007.

[3] J. Ye, K.-C. Leung, and V. O. Li, “Optimal delay control for combating bufferbloat in the Internet,” in Proc. 2016 IEEE International Conference on Communication Systems (ICCS), pp. 1–6, 2016. [4] J. Gettys and K. Nichols, “Bufferbloat: Dark buffers in the internet,” Queue, vol. 9, no. 11, p. 40, 2011.

[5] R. Pan, P. Natarajan, C. Piglione, M. S. Prabhu, V. Subramanian, F. Baker, and B. VerSteeg, “PIE: A lightweight control scheme to address the bufferbloat problem,” in Proc. 2013 IEEE 14th International Conference on High Performance Switching and Routing (HPSR), pp. 148–155, 2013.

[6] V. Misra, W.-B. Gong, and D. Towsley, “Fluid-based analysis of a network of AQM routers supporting TCP flows with an application to RED,” in Proc. ACM SIGCOMM Computer Communication Review, vol. 30, no. 4, pp. 151–160, 2000.