This year marked a coalescence* around the 28 GH band for 5G fixed access services. A year ago at MWC 2016**, different millimeter wave bands (mmWave) were still in play including the 70 GHz range. I recall discussing the performance of those solutions with exhibitors. They all claimed that mmWave access works in non-line-of-sight conditions. I even recall one the demos had a node behind a wall to make that point. The wall turned out to be made of cloth!
In any case, the 28 GHz band has better coverage characteristics than the 39, 42, or 70 GHz bands – that’s physics. Still, the outdoor-to-indoor penetration is extremely poor, making it challenging to build a scalable business model.
In my experience, two complementary factors stand out to destroy the business case for wireless fixed access services:
1- Underestimating the number of sites required to provide the desired service level. This is a capex issue where unrealistic estimates of a technology performance leads to under-sized, under-provisioned network that fall short of customer expectations.
2- Low service reliability which increases churn and raises the cost of customer acquisition and retention.
mmWave solutions are attractive because of high data rate. But this is contingent on good coverage. Beamforming is used to mitigate the inherent coverage deficiency. Yet, even with beamforming it is impossible to overcome high wall penetration losses. This leads vendors and service providers to ask subscribers to deploy CPEs ‘at the window.’ The window, of course, has to face the base station. This opens a plethora of topics related to the deployment model, customer support and many operational questions that have high impact the business case.
Wall Penetration Loss [Source: Nokia]
Coverage performance of mmWave vs. 800 MHz [Source: Huawei]
The performance impact on the business case is what operators are assessing. Ongoing trials seek to explore the performance criteria and corresponding investment. Deployment, operational and service models are critical aspect of this exploration. But choices maybe limited. When a subscriber CPE needs to have a clear shot of the serving base station, we can expect a targeted deployment, i.e. truck rolls, or a blanket of low-utilization sites (small cells) – a model that failed, to date, due to different considerations.
The business case for fixed wireless is often shown to compare favorably and even exceed that of fiber. But such analysis misses a number of points. For instance, while the costliest expense in fiber networks is putting fiber in the ground, the capex can be amortized over 20+ years. In contrast, it is difficult to know in advance the opex for wireless networks which makes for the bulk of expenses.
When it comes to new technologies, a good dose of realism helps investors in accounting properly for the associated risks. Wishful thinking often overcomes realism in assessing new business plans and leads investors astray. One should not underestimate the complexity of mmWave even with the technology evolution we have seen over the past 10 years.
** For MWC 2017 overview click here.
* Candidate frequency bands for 5G:
USA FCC NPRM: 27.5-28.35, 37-40, 64-71 GHz
Japan: Candidate bands include 3.6-4.2, 4.4-4.9, 27.5-29.5 GHz and all the bands in WRC-19 AI 1.13
China: 3.4-3.6 GHz (for trial); other candidate bands include both sub-6GHz and 24.25~86 GHz
Korea: 26.5-29.5 GHz (for trial; 1 GHz bandwidth per operator); other bands may be decided later