Simulation And Refinement of Random Early Detection, Active Queue Management to Minimize Packet Drop Rate

Computer Science and Engineering

There are numerous issues in wireless sensor network including node deployment, energy consumption without losing accuracy, data reporting model, node/link heterogeneity, node deployment, fault tolerance, network dynamics etc. Proposed work to control the congestion using fuzzy rule set and refining this rule set with the help of neural network module. For this, I have used the previously used congestion control technique called Random Early Detection for Diffi-Serv (Differentiated Service Network). After refining the fuzzy based RED (Random Early Detection) through the use of neural module, I compared the performances of original RED (Already Implemented in NS2 [SHAL2000]), fuzzy RED (based on fuzzy rule set), neuro-fuzzy RED (refined rule set with the help of neural network module), through simulation results. These simulation results are based on two factors, first is packet delivery ratio vs time, second is packet drop ratio vs. time. Various graphs have been constructed through simulation for comparative study of these 3 implementations of RED (Random Early Detection). The RED implementation for Diffi-Serv defines different thresholds for each class. RED simply sets some minimum and maximum dropping thresholds in the router queues. If the buffer queue size exceeds the minimum threshold, RED starts randomly dropping packets based on a probability depending on the average queue length. If the buffer queue size exceeds the maximum threshold then every packet is dropped.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.