5G – Peak Complexity (and how to make things simpler)

In a network that has been fragmented functionally, has an architecture that exposes network services across the user, control, policy and management planes, has vast topological options and can then be sliced so that all of these options can be repeated over and over again, how can we expect to deploy, operate and monetise 5G-era services? We are going to need assistance – and it is not necessarily coming in the form of humans. Much is made of Machine Learning, but on the road to an automated, self-managing, adaptive network, there are going to be mistakes, so we have a lot of learning to do too.