I think the article gets network slicing wrong. Using network slicing instead of the traffic shaping that the mobile edge router probably supports is a strange idea. The showcase network slicing use case was public safety comms. Network slicing is AFAIK not much used because it clashes with roaming. The article describes network slicing as reserving spectrum for certain apps. That's not how it works. It reserves capacity.
Still, zero-rating and traffic shaping should not be used to favor apps, especially not on a pay to play basis, for all the same net neutrality arguments as ever.
So surprised I had to scroll too far for this reply. I actually work for one of the major US carriers. My job is literally to figure out how to apply the technical capabilities of 5G to solve business problems. NONE of the US carriers have figured out how to actually deliver network slicing beyond, say, reserving capacity for first responders. And, as you say, it’s about capacity, not speed per se. We want to make sure that, say, an AGV can offload kinematics to the MEC and navigate in real time in dynamic environments. The poster child for network slicing is the surgeon doing telesurgery over a 5G network (But that’s likely to remain a poster child). We’re figuring out how to provide network slices for autonomous vehicles, mobile teleoperation, etc., in all use cases we’re examining it because something BAD could happen absent guaranteed capacity. I have never ever heard anyone talk about using network slicing for QoS for consumer apps.
This is correct. Slicing can offer significant performance gains in certain situations. For example, lower latency when certain users need it while not overburdening the network by having to give that to every user.
Still, zero-rating and traffic shaping should not be used to favor apps, especially not on a pay to play basis, for all the same net neutrality arguments as ever.