Will 5G "Succeed," and "How?"

It is probably easier to overstate 5G importance than to understate it. Some proponents might argue that 5G is qualitatively different from 4G in ways that are outcomes relevant. The issue is that many of the innovations using 5G actually are based on something else: internet of things or edge computing or private networks, for example. 

source: GSMA 

Even network slicing, the ability to create virtual private networks, end to end, provides value that can be supplied in other ways. If lower latency is required, then edge computing is a substitute. If higher throughput is required, then both local and wide area private networks are a substitute.  

It also is possible to understate 5G value. To the extent that 5G enables latency reduction of 10 times, plus bandwidth increases of 10 times or more, that value is analogous to improvements of fixed network bandwidth from 100 Mbps to 1,000 Mbps. 

There never was any question but that 5G “would succeed.” But what questions do exist have to do with enabling value. Will 5G lead to new use cases with important use case implications, business model outcomes or other innovations that change outcomes in important ways. 

Mobile operators have been hopeful of 5G revenue boosts. That has happened. But the larger context is that all ISPs must supply ever-more bandwidth over time, often with very-little incremental revenue. 

In that sense, 5G was essential, simply to maintain the existing business case for mobile internet value. Incremental revenue, value and use cases are a plus.

Big Connectivity Problems Sometimes Get Solved, and Nobody Notices

Sometimes big problems get solved and we hardly recognize that has happened. Consider the problem of providing voice communications to all the people of the planet. Fifty years ago, that was a big problem, as “half the world’s people had never made a phone call.” 

These days, the availability of networks supporting that use case is nearly ubiquitous. According to GSMA figures, voice access is virtually complete. “No network available” is rarely a barrier to communications usage. 

Looking at mobile internet access, we are close to 100-percent coverage as well. Only about five percent of people cannot access the internet using a mobile network. Something in excess of 55 percent of people now connect to the internet using their mobile devices. 

source: GSMA 

Less than 40 percent of people do not yet use the mobile internet. It still can be argued that mobile internet does not offer the speeds of fixed networks, and that the cost per bit on a mobile network is higher than on a fixed network. 

Still, it remains true that for most people on earth, mobile networks are the way they use the internet. That trend has been in place for a decade or more. While home broadband subscriptions using fixed networks continue to grow slowly, mobile broadband is growing much faster. In 2021, mobile internet subscriptions outnumbered fixed connections almost seven to one. 

source: Ericsson  

So while there still are performance gaps between mobile and fixed networks, the “people cannot communicate or use the internet” problem that so perplexed policymakers fifty years ago now is largely solved. Performance and cost need to improve, and will keep improving. But “access” per se is no longer a major problem. Instead, the issue is the “quality” and “cost” of connecting. 

5G: Don't Overplay it; Don't Underplay

The good news about any new technology is that if and when it becomes ubiquitous, we stop having to convince people it is a good thing. On the other hand, early expectations also tend to be inflated. So 5G "is a good thing," as gigabit home broadband is better than 100 Mbps. 

5G will allow for easier creation of customized virtual private networks for guaranteed latency or bandwidth. But there will be other ways to achieve those values, such as edge computing or private networks. 

The better way of framing 5G is whether mobile internet access is a good thing; whether remote sensors are a good thing; whether high-performance computing is a good thing. 

But it is like any other form of digital infrastructure: what matters is why you might be able to do with it. 

In Some Cases, 5G Value Propositions are Relatively Weak

Whether you are examining consumer use of 5G or business use of 5G, the question “is it better than 4G; how much better; and for what use cases?” is not as easy to answer as you might think. 

“5G is faster” is the baseline value proposition most consumers will appreciate. Lower latency compared to 4G will not be perceptible for most use cases. That might well be a key attraction, however, for some enterprise use cases, where 5G might perform significantly better than other alternative platforms. 

Still, “value versus cost” is almost always relevant for any product, save those instances where a particular problem can only be solved in a few ways, all the ways are costly, and the problem must be solved despite the cost. 

The consumer version of that value proposition is something like :I have a new $1,000 phone and it uses 5G. I can get a 5G service plan for not much more than a 4G plan. Even if 5G performance is only incrementally better, why not use it?”

That might be the case even in instances where the amount of time the user would be connected to, and using, the mobile network for internet access is quite limited. 

Consider a consumer user working from home on a full-time basis, and not required to travel extensively as part of the job. In such cases, the vast majority of internet connect time will use Wi-Fi. On a gigabit home broadband connection, that means a 5G smartphone might well get 500 Mbps to 600 Mbps on a routine basis.

In other words, 5G will provide the most value for a 5G customer who is out and about more than is typical; cannot often connect to Wi-Fi often and uses devices in areas with higher-speed 5G and/or relatively lower-speed 4G. 

In some cases, 5G might appeal to users who want to replace home broadband with mobile-only usage. For most consumers, though, connecting to Wi-Fi instead of 5G will have some advantages, even when those users are on 5G unlimited usage plans. 

source: Celltech 

Not many 5G connections can match that. So then the value proposition changes: “how valuable is faster 5G when it represents two percent of my total connection time; and when much of that connection time is when I am in transit, and I am not using the device actively?”

In such instances, even faster 5G is not going to deliver very much incremental value over 4G. 

Proponents of private 5G networks for enterprise use cases, including manufacturing, might find it easier to justify a new investment in a private 5G network, though. That would be especially true when 5G ultra-low latency actually provides value for process control.

5G could make sense when the device density is very high, as well, when coverage of larger areas is required (airports, seaports, campuses) or when both process control of machin

Which 5G Network is Best: CBRS, DAS or Millimeter Wave?

A study of deployment costs for 5G private networks, comparing use of licensed Citizens Broadband Radio Service (CBRS), millimeter wave (using a mobile operator’s spectrum) and a distributed antenna system approach--with spectrum assets supplied by a mobile operator-- suggests millimeter wave networks are most affordable when high capacity is required.

When lower capacity will suffice, CBRS is the lowest-cost option. The study, conducted by Mobile Experts, is of a private network supporting robots in a factory. 

The lower cost of the millimeter wave option in a high-capacity scenario is the small number of radios (one) compared to the CBRS and DAS alternatives. 

source: Mobile Experts 

But CBRS turns out to be the lowest-cost option when less capacity is required. In either scenario--high or low capacity--DAS was the most-expensive option.

Metaverse Will Shape Mobile Networks, Ericsson Says

It seems intuitive that coming extended reality apps and environments including metaverse will require network innovations related to capacity and latency. 

“The move from traditional 2D media to advanced immersive media services increases the informational load, due to the multiplicity of media streams and the increased media quality requirements. It puts high pressure on processing and transmission bit rates across the whole communication chain asymmetrically depending on how the XR use case is implemented – that is, it can impact the uplink, the downlink or a combination of both,” says Ericsson CTO Erik Ekkuden.   

“For instance, device spatial-mapping compute offload (to edge/cloud) will result in a more symmetric traffic load in the downlink and uplink compared with mobile broadband (MBB) traffic, which is mainly heavy downlink traffic,” says Ericsson. 

To guarantee quality of experience, “stringent bounded latency requirements are needed when device compute is offloaded to the edge and the cloud,” Ericsson notes. Some of that improvement can come from on-device processing. 

Network slicing and other latency improvements in radio access networks also will help.

source: Ericsson

Telcos have Practiced Demand Management Before. Maybe it Is Time to Do So, Again

Cell tower electrical consumption is related to content usage by mobile subscribers, it stands to reason. But just as data center operators seek ways to reduce power consumption, will mobile operators soon have to do the same?

According to Sandvine, watching video in 720p on phones uses 1/20th of the energy resources of a 4K video. Assume the average cell site today uses about 5kW of power. Further assume power consumption of about 0.4 kW per square kilometer. 

Assume a typical mobile operator uses about half that amount of power in any 24-hour period, allowing for power savings at night. 

That would come to about 462000 MWh of usage, according to Sandvine. Using 2021 electricity pricing ($0.19/kWh), that would cost an operator about $86 million, or about four percent of revenues, Sandvine estimates.

Part of the solution is demand management, says Sandvine. In other industries, such as the electrical power industry, there often are differential charges for usage at peak hours, for example.

Mobile operators used to do this themselves, as when they offered discounting long distance calling after 7 p.m. and before the morning business hours, as well as discounting calling on weekends. In all those periods, network demand was quite a bit lower. 

It is not clear how well similar demand shaping might work in the internet access business, but it might eventually be necessary. 

source: Sandvine

 

Data consumption presents similar issues. But it is a complicated matter. Internet service providers have customers. Some of those customers are heavier users. Only sometimes do service plans correlate usage and cost very closely. 

And that is but part of the issue. Customer data demand is driven, in part, by the applications they wish to use. And much of that demand comes from a relative handful of content and application providers who are not, strictly speaking, “customers.” 

To be sure, differential pricing policies, as well as measures to shift demand, have been used by telcos in the past. Business customers paid higher rates. International long distance drove profit margins for firms whose consumer customers sometimes were negative margin, zero margin or low margin profit contributors. 

But all those policies and prices were levied on customers. What is new is the effort by some regulators to shift burdens from customers to content and app providers whose products customers want to use. 

It is akin to taxing suppliers of washing machines, air conditioning and refrigerators because they represent a disproportionate share of network load at peak summer hours, instead of charging the actual customers who consume electricity. 

It is fair to argue whether that makes good sense. Perhaps less disruptive or questionable policies, such as a return to forms of demand management by ISPs themselves, should be tried first. Pricing policies can incentivize economical use of network resources, instead of disincentivizing usage. 

Telcos have done this before. Perhaps they should try it again. And perhaps regulators should allow them to do so, instead of instituting radically-different policies that might be questionable in terms of fairness, equity or logic. 

Perhaps ISPs should simply try what they have done before: incentivize customers to use network resources rationally. Right now, we do not see that.

MVNO Model Spreads to Fixed Networks Business

The mobile virtual network operator business model and fixed network structural separation are analogous ways of enabling retail services in either mobile or fixed network domains. In each case, the supplier of retail services operates using wholesale capacity and services purchased from another underlying network owner. 

The advantage for retailers is lower capital investment and faster time to market, typically paired with efforts to achieve lower operating costs. 

But wider use of the asset-light model is spreading in fixed network access and data center segments of the business. More service providers are seeing either joint ventures or asset sales as ways of speeding deployment of expensive fiber-to-premise networks operating in competitive markets. 

Investors, on the other hand, have come to view digital infrastructure as they historically have viewed other real estate assets including hotels, airports, seaports, power utilities or transportation assets. 

Also, more data center assets are moving to private equity or other institutional investor ownership for similar reasons. They are viewed as stable cash-producing assets with some scarcity value. 

For private equity investors, there typically also is a belief that the operating models can be improved to drive higher revenue and profits. 

One recent example of that trend is the reorganization plan adopted by Oi. 

Brazilian operator Oi, which had entered bankruptcy in 2016, is moving ahead with a slimmed-down operating model roughly analogous to that of a mobile virtual network operator, which leases wholesale capacity and services from a facilities-based service provider. 

To make that shift, and shed debt, Oi has shed its mobile assets, cell towers, data centers, video entertainment operations. It also is structurally separating its fixed network infrastructure operations from its retail fixed network operations, but will retain a minority stake in the infrastructure assets supplier. 

Oi’s mobile assets were sold to TIM (Telecom Italia) as well as Brazilian mobile operators Vivo and Claro Americas. The mobile cell tower assets were sold to Highline, a unit of DigitalBridge, which invests in digital  infrastructure. 

source: Oi 

The data center assets were sold to Brazil-based private equity firm Piemonte Holding. The video subscription assets were offloaded to Sky Brasil. 

A controlling stake in its fiber infrastructure business V.tal was sold to a group of investors headed by Globenet Cabos Submarinos and BTG.

After the structural separation of the fixed network business, Oi will continue to hole a minority stake in the infrastructure wholesale business, and operate its retail business--anchored by internet access--as a retail customer of the infrastructure business.

It is akin to the model used in the United Kingdom, Australia, New Zealand and Singapore, for example. 

The moves also illustrate the shift of ownership of digital infrastructure assets from service providers to private equity and other institutional investors that began decades ago with decisions by mobile operators to offload ownership of cell towers. 

Since then, a wider range of assets have begun to shift, including local access networks and data centers. 

In essence, a wider range of physical infrastructure assets are considered for disposal, to enable a lighter-asset business model that has been proposed by many observers for more than a decade.

Sri Lanka Economic Collapse Now Leads to 20% Higher Communications Costs

Apparently in response to an on-going economic crisis, the Telecommunications Regulatory Commission of Sri Lanka has raised mobile and fixed network prices 20 percent, starting September 5, 2022.

The virtually-predictable outcome is that subscriptions and usage will decline, as the shock of higher prices causes demand to fall. 

Charges for pay-TV services will be raised by 25 percent. The unusual move apparently is a way of compensating for a deteriorating currency triggered by a near-collapse of the Sri Lanka economy. 

Facing its worst economic crisis ever,  some say “everything that could go wrong with the economy has: Sri Lanka faces budget and current account deficits, hyperinflation, a devalued currency and a huge sovereign debt that it can no longer pay.”

source: Telecommunications Regulatory Commission of Sri Lanka 

It might be argued that a country with a food crisis, debt repayment crisis, hyper-inflation and a devalued currency has bigger problems than the price of using mobile and fixed communications. But it also is true that hyperinflation drives up prices of every good and service. 

Most service providers would prefer higher prices. But few would want higher prices driven by runaway inflation and economic collapse.

Are Access Providers "Doomed?"

One never hears a sales executive predict the death of the firms that constitute the market. That simply would not make sense, any more than it would for a supplier to repeatedly insult such potential customers. 

That does not mean markets do not evolve, and that some markets functionally cease to exist. But one simply will not hear vendors say so in public. Nor can a public company executive afford to say anything that does not accentuate the positive. 

Matters arguably are different for industry equity analysts or researchers. There, repeatedly incorrect and overly-robust assessments will damage the credibility of the assessor. 

So consider the nuance of an assessment of the industry's future by a supplier to that industry. 

“There are these notions of the death of the CSP and I think it’s silly,” says Stephen Rose, IBM’s global General Manager for Telco, Media, Entertainment and Distribution Vertical and Distribution.

The essential meaning of that statement is that the connectivity function always will exist, just as the computation and storage function will always exist. 

What is not said: how well will such connectivity providers fare in the future? Which entities will provide the function, and what will their business models look like? What market structures will evolve? What firms might wind up as the market leaders?

All of those questions are reasonable, even if one argues “connectivity providers always will exist.” Of course they will. The function is essential and foundational. But all other related questions might well be quite a bit more open and variable than can be admitted. 

People and businesses need internet access as they need computation and storage. But how they source it could change. Over time, some former “revenue drivers” become “features” of a business model. 

Over time, the leading suppliers of functionality change. And that could dramatically change the fortunes of firms and whole industries.

Better Broadband is Mostly Going to Be Used for Media Consumption

By some estimates, mobile internet access constitutes about 54 percent of total end user data consumption. About 63 percent of YouTube viewership happens on mobile devices. 

The majority of social app usage also happens on mobile devices. 

App	Mobile Usage Share (%)
Twitter	80
Facebook	81
LinkedIn	57
YouTube	63
Reddit	72

source: Zippia 

In the U.S. market, which historically features heavier PC usage than most other countries, relatively little online time is spent on “work,” as this data suggests. 

source: Martech.org 

More recent data shows the same pattern. In January 2022, of roughly seven hours spent per day using the internet, nearly six hours was spent consuming video or using social media. Mobile usage arguably is more weighted to media consumption.  

source: Hootsuite 

All that is interesting simply because the argument for quality and universal broadband is the purported productivity and educational benefit. But all that is “necessary but not sufficient” for productivity increases. 

All studies show that most internet usage and mobile device usage is for entertainment video and social media. True, there might be some ancillary productivity benefit there, but it is likely very little. 

That is not an argument against extending quality broadband. It is simply to recognize that the primary end user activity is entertainment and social media. 

----------------------