The enterprise network – AI’s most important frontier

November 11, 2021

When people think of everyday applications for artificial intelligence, they are likely to reach for examples like self-driving vehicles, digital assistants in the home or perhaps automated production lines in factories. They might well consider the use that banks and retailers make of AI to better understand patterns of customer behaviour. Many would be stumped to explain what AI means in the context of a wide area enterprise-grade network. Yet that is arguably where much of today’s most ground-breaking AI research and development is happening. Let’s consider some of the transformative ways in which AI is driving the connectivity of tomorrow.

Network managers, it seems, are forever being asked to achieve more and more, often with less and less budget at their disposal. AI is really the only realistic way for them to reverse this situation. AI lets them take on big challenges like growing levels of cyber threat, a multitude of clouds to manage and staffing constraints without having to spend more.

AI can additionally help spot trouble in the network before it actually happens, and resolve network problems before humans are even able to recognize that a problem exists. Event correlation and root cause analysis combine with highly nuanced mining of data to identify the specific bit of a network where there might be a problem, and potentially isolate other parts of the network from risk. Given the interconnectedness of corporate networks with those of customers and suppliers, this is an important way to block off trouble before it becomes an embarrassment. AI makes existing networks run better and safer, but is also an essential prerequisite for any new greenfield developments. If you are launching a new network then you want Day 0 to 2+ operations running as smoothly as possible.

Ask any CIO what their biggest network headache is, and they are likely to say ‘complexity’. AI plays an increasingly important part in reducing the complexity of networks, especially as their reach extends beyond the obvious head office and branch locations to embrace a more fluid network edge. Once the CIO, or network manager, has got a grip on all this complexity by automating it, then they can afford to shift their focus toward more strategic and high-value tasks. The kind of data mining required to resolve the hundreds of tiny problems that bedevil networks is no longer a manual drag on resources.

But perhaps the most important aspect of networking that AI is transforming, thanks to the efforts of various leading edge vendor names, is inside the data centre. It is here where AI is controlling the movements of machine learning workloads in containers and virtual machines, making them as secure as they are manageable. Work in this area is creating a new sweet spot in hybrid cloud computing, one promising greater control, lowered costs and the highest performance. This might not be as glamorous as a Mercedes that drives itself, or as visible as a robot that oversees a production line, but it is arguably a lot more transformative in terms of how future economies will flourish.

The following event will explore these themes in much greater depth than I can manage in this blog.