Towards the Automation of Data Networks

Abstract

A wide variety of enterprises, corporations, communications service providers (CSPs), and experts have highlighted the difficulty of managing modern networks. These networks exhibit high-impact technological innovations, such as cloud computing, mobility, new traffic profiles, network functions virtualization (NFV), the Internet of things (IoT), Big Data, among others. Network automation is a methodology in which physical and virtual network devices are automatically configured, provisioned, managed, and tested using software. Large enterprises such as Cisco, Juniper, Red Hat, and VMWare offer proprietary solutions for network automation. Additionally, the number of tools assisting in network automation has recently increased. Taken together, these developments have changed the way administrators build and manage networks. In this regard, most large communications operators are now working and moving toward truly autonomous networks that will eventually require an intensive use of Artificial Intelligence (AI). Advances in the area show that three specific segments —CSPs, Cloud Providers, and Enterprises— are all at different stages of automation maturity. Over time, network automation is expected to reach smaller organizations as well. This paper presents a specialized, detailed and current technical study on the state of the art in network automation, highlighting the trends observed in information technology (IT) environments, enterprises and communications operators —which are closely involved in this technology—, and concludes with a discussion on automation tools.