What the Heck is 5G?

Every time a new “G” comes around, the engineers try to define what it means in terms of technology.   This worked fairly well for 2G, 3G, and 4G, where each generation involved a new way of scrambling the bits for highest spectral efficiency.

But 5G is different.   We’re getting close to physical limits on pure spectral efficiency, and more investment along those lines would be wasteful.      It’s worthwhile to take a moment and ask ourselves:  What do we really want to achieve with 5G?

The top OEMs are starting to publish some of these “radar” charts which illustrate how performance will improve between eight and ten different dimensions at the same time.   

The next big thing could be to improve single-user data throughput to 10Gbp, and to reduce latency below 1ms, and to improve energy efficiency by a factor of 100.    In some of these fantasies, there is a naïve idea that 5G will improve ALL of these dimensions at the same time.

Did anybody notice that cost is not on the list of things to be improved?    My view is that 5G should focus primarilyon cost reduction, and other factors such as throughput or latency are simply a way to get the job done. 

Steve Jobs isn’t here anymore, so there’s nobody to tell us what applications people want to use in 2025.    We don’t know.  But we do know one thing---Machines and people are going to communicate differently:  

§  Machines can be patient, they don’t mind transmitting twice if necessary, and they don’t move around a lot.   There are exceptions, but most applications should fit this description.

§  People move constantly and get annoyed with any glitches or delays.   People will charge a battery, and will pay for a device that they use every day. 

 Should we really expect one technology to satisfy both of these requirements?

In fact, the main distinction between machines and people may be this:   Much of M2M communications can succeed with current wireless transmission cost, but the terminal cost needs to be reduced.     For human communications, the cost per bit needs to come down but the terminal cost is okay.     Aren’t these two things fundamentally different?

At the end of the day, the main thing that IoT and human applications have in common is the need for cost reduction.   Let’s avoid getting carried away by dreams of magical technology.  Instead, let’s focus on cost-optimized solutions for two distinct markets.