Fixed wireless is key to justifying 5G investment

Verizon CEO Lowell McAdam stated in May 2016 that revenue generated from fixed wireless services alone justifies the investment required for a 5G network. The rest (by which he meant 5G mobile) “is going to be gravy”, and will probably have to wait until 2020.

Most people who have followed 5G will be familiar with the three use cases that are usually cited as the commercial justification for it:

Enhanced mobile broadband
Massive connections (that is, IoT)
Ultra-reliable networks.

Fixed wireless represents a fourth use case. Fixed wireless is likely to be the first to appear, and could be of greater commercial value than the other three combined for several years to come. However, the use of fixed wireless technology for 5G raises many techno-economic questions. This comment briefly addresses the main questions.

To overcome the significant path losses at higher frequencies and to achieve outdoor-to-indoor coverage, Verizon will have to employ higher-order MU-MIMO/beamforming, and use wide channels (for example, 100MHz). Bandwidth gains will come from the number of antennas used, not from bps/Hz efficiency gains.

What sort of operator will want to use 5G fixed wireless?

There is a prima facie business case for any operator that has not already invested in next-generation access (NGA) sub-loop infrastructure.

This includes large mobile-first operators - FTTC incumbents that have not upgraded their physical infrastructure beyond the cabinet and greenfield new entrants.

AT&T, BT and Google Fiber have all publicly indicated interest in what appears at face value to be the same principle, and we are aware that other major operators are interested.

Why use 5G fixed wireless and not FTTP?

Google Fiber is frustrated by the rate at which it can roll out gigabit services using FTTP/GPON. One of the two main attractions of 5G fixed wireless is the potential speed at which it can be rolled out. The operator will still have to deploy some level of FTTN, but will not have to install the final connections to subscribers or fiber in multiple dwelling units (MDUs).

The second main attraction is that the infrastructure built up makes 5G mobile a manageable incremental cost. Other fixed-NGA technologies (such as FTTP and G.fast) will not be able to match that.

Where would an operator roll it out?

MDUs and dense urban settings appear to be most promising. High frequencies mean short range. Moreover, MDUs are challenging for NGA players because they add another layer of complexity for planning. Contrary to popular belief, dense urban areas do not represent quick wins for FTTP operators, especially if they are in cities with old housing stock and infrastructure - although G.fast gets around this complication. If 5G cell sites are going to be reused for 5G mobile, then urban, with its high footfall, looks best.

This is quite unlike earlier business cases for fixed wireless, which emphasized the low entry costs of wireless for sparsely spread premises. Macro cell infrastructure and underutilized FDD spectrum were frequently shared with mobile. Erecting monopoles for dedicated fiber-like fixed wireless in rural areas is an expensive way of doing NGA, and there is probably little commercial use for 5G mobile in rural areas.

What density of cells will be required?

The answer to this depends on whether the goal is fiber-like or fiber-lite. Verizon has indicated that up to 1km radius is possible. In a cluttered, urban residential setting, this seems unlikely - academic papers indicate that

If, as early trials claim, the operator could get 2-Gbps cell capacity (that is, not theoretical peak capacity) shared among 50, then that is about half what is typical on a GPON, but this still leaves plenty of headroom.

How much fiber backhaul is needed? 200m intervals may require more than just metro fiber for backhaul.

For operators without dense fiber, wholesale Ethernet is expensive, but is under commercial pressure from competition, and many operators are willing to sell dark fiber. Indeed, regulators may oblige incumbents to sell dark fiber. Dedicated optical wavelengths will be another option, and cheaper if on NG-PON2. Self-backhauling 5G will be possible over short distances in the first aggregation hop, but operators will prefer fiber if it is already available.

Assuming a rough equivalence in the cost of the horizontal fiber network, 5G fixed wireless will probably be cheaper than FTTP (except where this is aerial), but cost more than G.fast. We cannot yet know how much base stations or remote radio units with large arrays will cost. Even if 5G fixed wireless turns out to be more expensive than wired alternatives, it may not matter in the long run because those cells can be reused for future 5G mobile.

How much 5G mobile ‘gravy’ will there be?

This is too big a question to answer here. However, the introduction of ever-higher speeds and capacity in 4.5G networks has diminishing returns. Investing in 5G just for enhanced mobile broadband looks like pedaling hard to stay still. 5G IoT faces challenges from LPWA and we need to be cautious about potential carrier revenue.

The ultra-reliable, high-performance use case looks as if it will take the longest to see the light of day, but is also perhaps the only use case that exploits the unique capabilities in the 5G performance targets.

There will still be two distinct fixed-mobile convergence investment models. 5G is at the heart of convergence, but this does not mean a convergence of investment models. Broadly speaking, two are emerging:

Outside-in: Wireless provides all access, but requires a huge investment in cell densification and fiber.
Inside-out: Access is a mix of fixed and wireless. The provision of 5G mobile x-haul is part of the business case for FTTP; 5G fixed wireless is, at most, complementary.

The winners will be determined not by technological superiority, but by capital efficiency.

Rupert Wood is lead analyst for Fixed Networks and Wireless Networks at Analysys Mason

This article was first published on Telecom Asia 5G Insights September 2016 edition