Congestion control algorithms for a new TCP/UDP router based on 2-D stability conditions

  • Authors:
  • Yang Xiao;Lei Wang;Jun Niu;Seok Woo;Kiseon Kim

  • Affiliations:
  • Institute of Information Science, Beijing Jiaotong University, Beijing, China;Institute of Information Science, Beijing Jiaotong University, Beijing, China;Institute of Information Science, Beijing Jiaotong University, Beijing, China;Dept. of Information and Communications, Gwangju Institute of Science and Technology, Gwangju, Republic of Korea;Dept. of Information and Communications, Gwangju Institute of Science and Technology, Gwangju, Republic of Korea

  • Venue:
  • WiCOM'09 Proceedings of the 5th International Conference on Wireless communications, networking and mobile computing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Classical network fluid model and RED algorithm are based on TCP flows in internet network, and it has not considered the UDP flows' effects in network. However, in real work of the network, the network link capacity is shared by TCP flows and UDP flows, UDP flows can occupy the original link capacity of TCP flow. Since there is no feedback control for UDP flows, the classical network fluid model and RED algorithm can not reflect and control the congestion of TCP/UDP networks. In this paper, we develop a router scheme with classifier, access control, separated TCP queue and UDP queue, and revised AQM algorithm. The proposed TCP/UDP router can be expressed by a linear time-delay system model; this paper derives some stability conditions with simple router parameters by utilizing the 2-D Laplace-Z transform technique. This paper proposes parameter setting for TCP/UDP routers for congestion control based on 2-D local stability conditions. A serial of simulation results verify that the proposed router can gain the effective congestion control in regards to the stability conditions. The performance comparison for different TCP/UDP network scenarios with the stability conditions is provided.