Performance analysis and modeling of congestion control algorithms based on active queue management

Abstract

We have considered various types of congestion control algorithms. Each congestion algorithm has its own advantages and it would vary from parameters to parameters. Random Early Detection (RED) is more focused on queue length and BLUE is care about loss of packets. In this paper we have found the problems with existing congestion control algorithms. We have tried to show their performance of RED, SFQ, and REM in terms of performance parameters i.e. delay, throughput, loss rate etc. for our considered network configurations. Similarly BLUE and Random Exponential Marking (REM) are more focused on packet loss and mismatch respectively, mismatch occurring in REM due to either input rate and link capacity or queue length and target. In order to restrict the rising packet loss rates caused by network traffic, active queue management technique such as REM has come into picture. Flow Random Early Drop (FRED) keeps state information based on instantaneous queue occupancy of a given flow. Stochastic Fair Queuing (SFQ) ensures fair access to network resources and prevents a busty flow from consuming more than its fair share. Stabilized RED (SRED) is another approach of detecting nonresponsive flows. In this paper, we proposed a model to calculate dropping probability and packet loss for Active Queue Management (AQM). At the last, we have shown a comparative analysis of the loss delay product (LDP) as a new parameter of performance measure obtained from simulation on ns2 for different AQM algorithms. It has been observed that performance parameters are varying according to the various congestion algorithms used in the simulation. RED achieved the best result in terms of the delay but in terms of throughput, loss ratio, and utilization REM shows the best results in this network configuration. But, RED performed best at low link capacity in terms of new measured parameter LDP.