Loading...

Dynamic Modeling and Congestion Control in Computer Networks

Kahe, Ghasem | 2013

758 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44474 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Jahangir, Amir Hossein; Ebrahim, Behrouz
  7. Abstract:
  8. Active queue management (AQM) is a key factor in congestion control and should provide appropriate feedback for flow control in traffic sources to overcome the congestion problem. Beside providing congestion control, achieving predictable queuing delay, maximizing link utilization, simplicity and robustness are also the main objectives of an AQM controller. We propose in this thesis an improved queue dynamic model while incorporating the packet drop probability as well. The proposed model is evaluated using ns2 platform. By applying the improved model, a new compensated proportionalintegral- derivative (PID) AQM controller is developed for TCP network. The time-varying nature of the network dynamics and the complex process of retuning the current AQM algorithms for different operating points necessitate the development of a new AQM algorithm. Since the non-minimum phase characteristics caused from the Padé approximation of the network delay, restricts the direct application of control methods, we propose a compensated PID controller based on a new control strategy addressing the phase-lag and restrictions caused by the delay. A parameter-varying dynamic compensator, which operates on tracking error and internal dynamics, is proposed not only to capture the unstable internal dynamics but also to reduce the effect of uncertainties caused from unresponsive flows. The proposed dynamic compensator is then used to design a PID AQM controller whose gains are obtained from the state-space representation of the system with no further gain tuning requirements. Most of AQM algorithms are designed for a nominal operating point and re-tuning them for a new operating point demands complex and time-consuming analysis. Moreover, network parameters are time-varying and cause the operating point to frequently violate the robustness bounds of the controller which necessitate re-tuning the controller gains to meet the desired control objectives. The self-tuning structure of the proposed controller allow us to extend it to a self-tuning PID controller through network parameter estimation. Traffic load, network delay and the bottleneck link capacity are the time-varying network parameters whose effects should be compensated by the controller gains adaptation. The network parameters are estimated from measurements made locally at the congested router. The packet-level simulations using ns2 show the out-performance of the developed controller for both queuing delay stability and resource utilization. The improved underlying model leads also to the faster response of the controller. Simplicity, low computational cost, self-tuning structure and yet considerable improvement in performance are exclusive features of the proposed AQM for the edge or core routers
  9. Keywords:
  10. Computer Networks ; Internet ; Dynamics Models ; Congestion Control

 Digital Object List

 Bookmark

...see more