Networking professionals increasingly believe machine learning is a critical component of network management, a new survey suggests. That belief stems, at least in part, from the growing adoption of multi-cloud environments, according to network analytics company Kentik.
To find out what networking experts think about the use of AI and automation, Kentik surveyed 388 attendees at the Cisco Live US conference held in San Diego, Calif. last month. Half of the respondents were networking engineers, while other respondents included SVPs and VPs, directors, managers, architects and developers. They represented a range of industries, including education, energy, finance, government, health care and technology.
As many as 65% of respondents said that machine learning is now “extremely important” or “very important” for network management. That’s up from 45% in Kentik’s 2018 survey. The sentiment was consistent among executive-level respondents and those with technical titles.
Kentik linked the interest in machine learning for network management to the growing use of cloud environments. As many as 76% of survey respondents indicated they were using cloud services, and of those respondents, nearly half (47% ) were using a multi-cloud strategy.
“Humans and manual processes can no longer keep pace with network innovation, evolution, complexity, and change,” Jim Frey, VP of strategic alliances at Kentik, said in a statement. “That’s why we’re hearing more about self-driving networks, self-healing networks, intent-based networking, and other concepts.”
Indeed, during the Cisco Live conference, Cisco announced a series of software enhancements designed to put AI and machine learning deeper into the network. Kentik, meanwhile, offers an AIOps platform for network professionals.
Among survey respondents, those in the energy sector showed the most interest in machine learning, with 75% calling it “extremely important” or “very important” for network management. Not a single person in this vertical considered machine learning to be “not at all important.”
Respondents from the health care sector showed the least interest, with just 41% calling machine learning “extremely important” or “very important.”
Outside of the technology industry, the energy sector also proved to be the most prepared for full automation. 30% of energy sector respondents said their organization is “extremely prepared” or “very prepared” for full automation. In the health care industry, just 3% of respondents reported that their organization is “very prepared.”
The survey also showed that automation is largely deployed for network configuration, with 53% of respondents citing that particular use case. Policy management was the second-most automated process, cited by 40% of respondents.