In this thesis, a study is performed to find the effect of applications on resource consumption in computer networks and how to make use of available technologies such as predictive analytics, machine learning and business intelligence to predict if an application can degrade the network performance or consume computer system resources. In recent years, having a healthy computer system and the network is essential for continuity of business. The study focusses on analyzing the performance metrics collected from real networks using scripts and available programs created specifically for monitoring applications and network in real-time.
This work has significant importance because monitoring real-time performance doesn’t give accurate or concise information about the reasons behind any degradation in network or application performance. On the other hand, analyzing those performance metrics over a certain period and find a correlation between metrics and applications gives much more relevant information about the root cause of problems.
The findings proved that there is a correlation between certain performance metrics, besides correlation found between metrics and applications which conclude the study objectives. The benefits of this study could be seen in analyzing complex networks where there is uncertainty in determining the root cause of a problem in applications or networks.
Library of Congress Subject Headings
Computer networks--Reliability; Reliability (Engineering)--Mathematical models; Machine learning; Data mining
Telecommunications Engineering Technology (MS)
Department, Program, or Center
Electrical, Computer and Telecommunications Engineering Technology (CET)
William P. Johnson
Elmasry, Mohamed, "Predict Network, Application Performance Using Machine Learning and Predictive Analytics" (2019). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus