Advantages of Machine Learning in Networking-Monitoring Systems to Size Network Appliances and Identify Incongruences in Data Networks

Bustamante A.J.; Ghimire N.; Sanghavi P.R.; Pokharel A.; Irekponor V.E. · Lecture Notes in Networks and Systems (2022) · DOI: 10.1007/978-3-030-96147-3_15

Resumen

This paper shows two potential uses that Machine Learning makes possible on Data Networks, multiple algorithms have been used to predict future behaviors about common components in data networks to take actions base on the predicted results, the first advantage identified is for sizing resources in digital network appliances and identification of discordances in the traffic going through the networks which could be used for reducing security risks. The focus of the paper has been on monitoring systems working with protocols like SNMP, Netflow, and Syslog to recollect data and act base on that. Different algorithms have been proven beginning with linear algorithms up to neural networks to predict future behaviors and reduce security gaps as well as for sizing resources. Overall, in this paper it has been demonstrated clearly two benefits of using Machine Learning in Data Networks.