Abstract

With the increase of multicast data applications, research interests have been focused on the design of congestion control scheme for multicast communications. One challenge comes from the heterogeneous multicast receivers, especially those with large propagation delays, which means the feedbacks arriving at the source node are somewhat outdated and harmful to the control actions. To attack this problem, this paper describes a novel multicast congestion control scheme, which is based on the proportional plus neural network (PNN) predictive technique. The congestion controller is located at the multicast source and uses the explicit rate algorithm to regulate the transmission rate. This network-assisted property is different from the existed control schemes in that the multicast source is able to predict the available bandwidth at those receiver nodes for which the back control packets experience very long propagation delay and possibly cause irresponsiveness of a multicast flow. The proposed control scheme is more responsive to the network status thus the rate adaptation can be made in a timely manner for the sender to react to the network congestion quickly. Simulation results demonstrate the efficiency of the proposed scheme in terms of fast response, high link utilization and stable buffer occupancy.

Keywords: BP neural network; Explicit rate; Multicast congestion control; Proportional controller; Rate-based congestion control

Article Outline

1. Introduction

2. Congestion control model

3. The Algorithm

4. Implementation of P and PNN controller

4.1. P controller

4.2. The algorithm for stability analysis of BufferOccupancy

4.3. The PNN predictive control technique

4.3.1. The BP neural network architecture

4.3.2. Multi-step neural predictive technique

5. Simulation results

6. Conclusions and future work

References

Vitae