Software-Defined Networking Test Framework With Real Network Traffic Data Using Pcap

Date of Award

8-2021

Document Type

Thesis

Degree Name

Master of Science in Data Science

First Advisor

William Emmanuel S. Yu, PhD

Abstract

Software-Defined Networking (SDN) is an advancement in the field of computer networking. Previous studies have built SDN frameworks to test the performance of Quality of Service (QoS) mechanisms more easily than tradi- tional networks. A limitation of previous work in QoS and other use cases is the use of synthetic traffic. This research aims to extend the previous study through the use of real-world network traffic data, packet captures, in achieving more ac- curate and realistic performance tests. This was done through the use of PCAP file replay and minor framework modifications. The results of the PCAP file tests show that the performance comparison of leaf-enforced and core-enforced algorithms are consistent as those in previous studies. The results show that in Basic Class-Based Queuing (CBQ) 53.33% of 30 trials showed leaf advantage, in Source CBQ 70% of 30 trials showed leaf advantage, in Destination CBQ 100% of 30 trials showed leaf advantage, and in Source-Destination CBQ 70% of 30 trials showed leaf advantage. These results correlate with the results of the previous experiments using synthetic data.

Link to Full Text

Share

COinS