Approaching zero: Taking 5G latency to the limit.

This is a dramatic example, but one that will be possible with 5G. And at a time like this, latency would be a life-or-death issue.

Many new 5G use cases will require ultra-reliable low-latency communications (URLLC) for smart factories, smart grids, and more.

Network latency is typically counted in milliseconds (ms), which is one thousandth (.001) of a second. It can be measured as either round trip time (RTT), which is the time it takes a packet to get from the client to the server and back, or as the time to first byte (TTFB), which is the time it takes the server to receive the first byte of data from the client. We’ll use TTFB in this article.

Typical latency for a 4G network is between 30 and 50 ms. Part of the promise of 5G—a key feature that will make it transformative—is significantly reduced latency.

In the future, users will be able to experience latency lower than 10 ms with fully developed 5G. But we’re not there yet.

How low is low enough? How close to zero do we need to get to realize the full potential of 5G for use cases like autonomous vehicle fleets, the tactile internet, and augmented reality for tasks like remote building inspection or virtual training?

It depends on the application—and for those with a human interface, it also depends on which of our senses we’re using.

Humans can perceive a lag as short as 13 ms for visual stimuli. For a voice call, a sub-50 ms lag makes the call feel natural; when you get up around 150 ms, communication becomes untenable.

For other 5G low-latency applications—like when you need actual physical, haptic feedback—a 50 ms lag is an eternity. The classic, if overused, example: remote surgery. Could a heart surgeon remotely control a robot to conduct an emergency surgery? Could she actually feel what’s going on and respond instantly to changes?

That demands feedback between the robot and her brain that feels as fast as the feedback between her hands and her brain. Low latency is critical to achieve a sensation that feels to the surgeon like instantaneous feedback.

The impact of network latency can be easier to understand when applied to a familiar use case, like remote control. In the 5G era, this includes remotely-driven vehicles, where a driver in one location depends on low network latency to smoothly navigate a vehicle in another location. Because human reaction times vary, the difference between 100ms and 20ms network latency could mean the difference in experiencing or avoiding a collision.

There are two ways 5G can reduce latency: in the radio access network (RAN) and via multi-access edge compute (MEC).

The evolution of 5G is happening in phases, and at first, users are connected to both 5G and 4G. This is called non-standalone 5G. 

As 5G advances, non-standalone 5G is being gradually replaced by standalone 5G. The standalone version is pure 5G and is not shared with 4G at all.

The standalone 5G radio interface introduces flexibility to treat different traffic types in different ways according to priority, to speed more critical data that requires lower latency through the radio faster. For example, in a manufacturing company, the network could prioritize communications between machines and robots on the factory floor, which require lower latency than, say, HVAC sensors.

Here’s another way 5G can help reduce latency. In a 4G world, information has to travel miles farther to reach you. All those extra miles mean extra milliseconds—and extra latency.

With 5G, information won’t have to travel nearly as far when we push the application closer to the user via multi-access edge compute, or MEC.

Because we’ve lived in a high-latency world for so long, businesses don’t really think about latency—until they run up against a barrier. Instead, they build applications with latency limitations in mind.

In a low-latency economy, businesses will start to develop applications and use cases that seize this new opportunity.

A key theme will be enabling remote and autonomous devices to operate in near-real time without paying the latency penalty.

• Smart factories—that react in milliseconds to quality issues or changes on the line.
• Remote construction—guiding robots around unsafe or difficult-to-reach sites.
• Massive, multi-user VR experiences—from group simulations to “walk-through” holograms.
• Autonomous vehicle fleets—optimizing routes in near-real time.

These are the kinds of 5G low latency use cases that only become possible when you bring the core close to the edge to seriously reduce latency.

Here’s a closer look at a few low-latency 5G network applications.

Automated construction monitoring
In a low-latency world, AI and remote-controlled robots can help scale up specialized skills. Companies can operate highly intelligent inspection robots to monitor job progress and keep multi-million-dollar projects on track.

Industrial automation
A low-latency network will be critical infrastructure for the next wave of manufacturing innovation. As the Industrial Internet of Things builds sensors into every part, supply, machine, process, and warehouse location, AI-powered algorithms will optimize real-time manufacturing in ways no human could ever even attempt.

With all machines and robots wirelessly connected, you’ll be able to reconfigure entire production lines on the fly, improving your whole process to meet new needs.

The tactile internet
When virtual reality can be live streamed with haptic feedback, the potential applications increase dramatically. Tennis coaches will see exactly what their players see and feel the ball on the racket. An operator remotely manipulating a robotic hand will feel the fingers’ grip on an object.

And for each of these new experiences, there’s a business opportunity waiting to be seized. This is the promise of low-latency 5G.