Performance Benchmarking & Data Analytics for the Comparison of Multiple Data-Centre Network Topologies
Final project report for lab course "Data Science & Machine Learning"
TUMunich, Chair of Communication Networks (LKN), summer 2021
In the following, an analysis is performed on the data collected while testing and benchmarking a hybrid network topology called Cerberus. This network topology enables more efficient data transfer in data centres. The fundamental analysis requires large amounts of data to be visited and aggregated to gain insight into the performance of three types of network topologies. The main results were that the input distribution type could be verified as uniform and there are significant differences in completion times between topology configurations.
Download Project Report here: CerberusPerformanceAnalytics.pdf