A forecast was recently published, predicting “100 million 5G subscribers by 2025”. Give me a break. Come on people, 5G is not defined yet.
We don’t know what the use cases are.
We don’t know what kind of terminals will be used.
We don’t know if “subscribers” is a relevant metric for 5G.
We don’t have any standards for new technology, with only the first study items underway now.
And here’s the real problem: We don’t know what the business model is for 5G.
What possible basis do we have to estimate the number of “subscribers” on a given date?
Here’s a more rational perspective on 5G: As new applications with compelling business models emerge, technology will adapt to address the business model. Instead of wild guesses, let’s spend our time looking at whether new mobile apps are viable.
Look at the PC industry. Old folks remember the days of the 386, 486, and Pentium processors, when clock speed doubled or tripled in each generation. That trend changed when it became pointless to have higher clock speed, and when the applications demanded change in a different parameter. The microprocessors evolved into multi-core processors to satisfy new requirements.
There’s a lot of 5G interest within the mobile industry, because everybody has an R&D lab and a desire to be ready for the next big thing. In conferences, people are talking about a “grab bag” of different technologies such as Massive MIMO, machine-type communications, network virtualization, full duplex, and millimeter-wave bands. Basically, if it hasn’t been adopted as a part of 4G, somebody is calling it 5G. But most of this is driven by what the engineers can do, not by a new revenue stream. That worked okay for 2G, 3G, and 4G, just like the trends during the 286/386/486 era.
After all, the world has reached a high level of penetration for handsets, and ARPU for 4G is no higher than it was for 3G in most places. Is there any purpose for a 5G handset that can do ten times the data speed if nobody collects more money for it? Most major industry players recognize this fundamental issue, and they’re beginning to focus on enterprise applications as a new source of revenue. Many enterprises will pay for bursty real-time data which is delivered 99.999% of the time. Not many enterprises will pay for 10 Gbps in a mobile environment. The focus is likely to shift from throughput to new metrics such as latency and availability.
The next steps will involve consensus-building, standardization, some basic research, spectrum allocation, and finally product development. The process will take several years. 3GPP will open standardization opportunities in Release 14 and 15. ETSI has just launched some working groups. Governments will coordinate in WRC-2015 and WRC-2019 to identify spectrum, and are likely to license new spectrum only after 2019.
All previous generations (2G, 3G, 4G) have been centered on development of a new radio air interface protocol. In my opinion, it’s unlikely that 5G will follow this pattern. Instead, 5G is much more likely to encompass a network-wide overhaul to streamline, virtualize, and reduce cost. There may be new modulation levels or variations in existing options for OFDM, but the radio interface will not be the centerpiece of 5G. It’s about moving the cloud closer to the user, which drives change in the end-to-end architecture.
In short, the prediction of 100 million 5G “subscribers” is totally spurious and meaningless.
Mobile Experts will begin our coverage of 5G this year, but we won’t be making predictions about specific technologies or numbers of subscribers. We will begin with a look at wireless business ideas that are not well supported by 2G/3G/4G/Wi-Fi today. For cool new ideas that need a change in latency, or ultra-high reliability, or 10 Gbps throughput, then we will examine the potential profitability of the business model. We can start to make forecasts in a few years, after the basics have been covered.
Comments
Post a Comment