An Improved Network Congestion Control Strategy Based on Active Source and Link Management

An Improved Network Congestion Control Strategy Based on Active Source and Link Management

Xushi Mei (Jinhua Polytechnic, Jinhua, China) and Xiaolong Chen (Jinhua Polytechnic, Jinhua, China)
DOI: 10.4018/IJMCMC.2020010104

Abstract

When a source end adopted FAST TCP to actively control the sending window, and the link end adopted active queue management algorithm, if the network parameters were improperly set, the system can be unstable. The relationship between the stability and network parameters was quantitatively analyzed. Aiming at the defect of network instability caused by radical window halving strategy, a new method based on stability analysis was proposed to make full use of the historical change information obtained by the source-end and to improve the window adjusting strategy of FAST TCP protocol according to the congestion status. Thus, the parameters of the active queue controller was modified indirectly to improve the stability of the system. Ns-2 simulation result verified the accuracy of the stability theory analysis and the effectiveness of the improvement strategy.
Article Preview
Top

1. Introduction

FAST TCP (Koo, Choi, & Lee, 2008; Wei, Jin, Low et al., 2006; Li et al., 2016; Zhang & Chen, 2016) was a new network transmission control protocol proposed by Steven H. Low of Network Lab in California institute of technology, which was mainly used for the next generation high performance network with high speed, long delay and large capacity. Based on the balanced design idea of the FAST TCP, the stability, throughput, fairness and fast response were better than TCP Reno, HSTCP and STCP (Wei, Jin, Low et al., 2006; Deng, Chen, & Zhang, 2004). However, the above good performance was assumed that the bottleneck link end had a large enough cache, and adopted tail-drop queue management algorithm (DropTail) (Samreen, 2018). There was no packet loss in the network (Koo, Choi, & Lee, 2008; Wei, Jin, Low et al., 2006; Chen & Chen, 2016; Zhang & Chen, 2016). However, the packet queue cache management (Liu, Zhong, He et al., 2018) of intermediate nodes (such as routers) generally adopted the stochastic early detection active queue management algorithm (Random Early Detection, RED) recommended by RFC2309. According to the average queue length in the cache, RED algorithm will actively drop some packets at a certain probability (Huang, Lin, & Ren, 2006; Tan, Zhang, Peng et al., 2006; Zhang, Tan, & Peng, 2009). By analyzing the code of FAST TCP protocol (CUBINLAB, University of Melbourne, Australia http://www.cubinlab.ee. unimelb.edu.au/ns2fasttcp/), like other network transport protocols, FAST TCP will halve the size of sender window when the source end received three duplicate acknowledgment frames (packet loss). When there was packet loss, the above FAST TCP protocol good performance, especially the stability of the network could not be guaranteed. The reason was that FAST TCP protocol and Reno protocol adopted different window adjustment strategies, so the RED parameter setting guided by TCP Reno/RED stability analysis results based on low-speed network cannot ensure the stability of FAST TCP/RED system in high-speed network (Zhang, Tan, & Peng, 2009; Prakash, Tripathi, Pal et al., 2018).

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 12: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2020): 3 Released, 1 Forthcoming
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing