Examining the round trip time and packet length effect on window size by using the Cuckoo search algorithm
Irregular sequences of inter-arrival times of packet(s) and packet lengths in a network session determine effective traffic performance. Crucial to this is the width of the sliding window. This study utilized raw data from network traffic and built a Neural Network (NN) model trained with the Cuckoo...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English English |
| Published: |
Praise Worthy Prize
2016
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/55853/ http://irep.iium.edu.my/55853/ http://irep.iium.edu.my/55853/ http://irep.iium.edu.my/55853/1/002-Adamu_def_19233_%20%281%29.pdf http://irep.iium.edu.my/55853/7/55853-Examining%20the%20round%20trip%20time%20and%20packet%20length%20effect%20on%20window%20size%20by%20using%20the%20Cuckoo%20search%20algorithm_SCOPUS.pdf |
| Summary: | Irregular sequences of inter-arrival times of packet(s) and packet lengths in a network session determine effective traffic performance. Crucial to this is the width of the sliding window. This study utilized raw data from network traffic and built a Neural Network (NN) model trained with the Cuckoo Search (CS) algorithm. Round trip time (RTT) and packet length were captured over several network sessions. They were used as input and their effects were evaluated on window size as the output. Experimental analysis was carried out in order to test the model with various partitioning levels of training and test data. The results of the experiments show that the proposed NN model trained with CS successfully converged without any form of oscillation; the minimum MSE was observed shortly after 100 cycles. The predicted window size and target window size fitted each other. This signifies that the training was successful based on the fitted values of the window size. Thus the proposed model trained with the CS algorithm provides a high convergence rate to the true global minimum and a better optimal solution. Therefore, the combination of CS and NN (CSNN) contributed to decision making on the allocation of window size in determining network flow problems and congestion control |
|---|