Title :
Joint TCP Congestion Control and CSMA Scheduling without Message Passing
Author :
Xin Wang ; Zhaoquan Li ; Jie Wu
Author_Institution :
Dept. of Commun. Sci. & Eng., Fudan Univ., Shanghai, China
Abstract :
In this paper, we consider joint congestion control and wireless-link scheduling design for TCP applications over the Internet with ad-hoc wireless links. Differing from the existing methods, the main idea of our approach is to develop non-standard window-based implicit primal-dual solver for the intended optimization problem. Then queueing delays are employed to decompose this solver into local algorithms that can be deployed and operated asynchronously at different layers of network nodes. Capitalizing on this approach, we put forth the QUIC-TCP congestion control algorithm, and a new queueing-delay based carrier sense multiple access (CSMA) scheduling scheme. The proposed schemes can be implemented asynchronously without message passing among network nodes. Confined to the design space of TCP and CSMA, they are readily deployed for practical Internet applications. Moreover, global convergence of the proposed joint design to optimal network equilibrium can be established using the Lyapunov method in an idealized network fluid model. Simulation results are provided to evaluate the proposed schemes in practical networks.
Keywords :
Internet; Lyapunov methods; ad hoc networks; carrier sense multiple access; optimisation; queueing theory; radio links; telecommunication congestion control; transport protocols; CSMA scheduling scheme; Internet; Lyapunov method; QUIC-TCP congestion control algorithm; TCP applications; ad-hoc wireless links; carrier sense multiple access; global convergence; network fluid model; network nodes; nonstandard window-based implicit primal-dual solver; optimal network equilibrium; optimization problem; queueing delays; wireless-link scheduling design; Ad hoc networks; Delays; Internet; Joints; Message passing; Multiaccess communication; CSMA scheduling; Lyapunov function; TCP congestion control; convex analysis; network fluid model;
Journal_Title :
Wireless Communications, IEEE Transactions on
DOI :
10.1109/TWC.2013.100313.130116