Applied AI, IoT are the Issues, Not 5G

Lots of well-informed, smart people believe the deployment of 5G access networks is something of a race for economic leadership.  It is a complicated matter, though, since many of those observers also believe 5G is intimately connected with reaping the benefits of artificial intelligence and the internet of things.

Some of us would position the matter the other way around: economies, industries and countries that learn to innovate with AI and IoT will reap rewards. 5G is mostly along for the ride. In other words, firms, industries and countries that learn how to use AI and IoT to boost productivity and create important new use cases with big economic impact will “win.”

source: World Economic Forum

“The rollout of 5G is expected to enable and widely disseminate technologies, such as: the Internet of Things, self-driving cars, autonomous drones, and Star Wars-inspired hologram phones,” says a study by the World Economic Forum. And that is the point: 5G is infrastructure for reaping the benefits of potential innovation in IoT and from applied artificial intelligence.

But it will take time and perhaps some luck to discover and create all those new use cases. Deployment speed, in other words, might help, or might not, if innovation actually requires not only time and experimentation, but other enablers as well.

To use an older analogy, deploying higher-speed internet access networks is a necessary, but not sufficient, driver of economic innovation. What really matters is the ability to wring economic benefit and innovation from the existence of those assets.

source: Nokia

Yes, for some use cases, the ability to sustainably deploy IoT systems might hinge on 5G-specific capabilities (ultra-low latency or device density per cell site). In other cases innovations can occur using 4G, Wi-Fi or other access methods, plus edge computing.

To use another analogy, there are many reasons why Facebook, Google, Apple, Amazon, Netflix, Alibaba, Tencent, Baidu or JD.com arise, and where they arise. Plentiful broadband access helps, but does not automatically lead to creation of sustainable, global platforms, services and apps.

For whatever reason, internet platforms with scale have developed mostly in the United States and China. There are scale reasons, to be sure: both countries have huge internal markets. Still, innovation in the internet apps and platforms arena does not happen everywhere, even when goog internet access is everywhere.

Nor will innovation happen everywhere good 5G is available. Other forces are at work, and must be at work, for global scale and leadership to emerge.

source: EC  

The so-called race is about the ability to harness the power of AI and internet of things in economically-important ways, not the race to deploy 5G access networks, as such.

Very Few Keys to Success in Mobile Business

The key to success in the mobile business comes down to quite a small number of products, providers, revenue sources and investment decisions. You might say present revenue boils down to market share, which is another way of describing scale or “number of subscriptions.”

You might argue that revenue growth, on the other hand, is dependent on sales of mobile internet access services.

For 5G, you might argue everything hinges on the business case, with new use cases, spectrum costs or infrastructure requirements being key inputs. The point is that in all these cases, only 20 percent of actions lead to 80 percent of results, in business, life or telecom.

It usually is surprising how often a Pareto distribution occurs in business or nature. Most underlying trends in any business follow a Pareto distribution, commonly known as the “80/20” rule, where 80 percent of results flow from some 20 percent of the instances. That applies for consumer manufacturer warranty claims, for example.

In the telecom and most other businesses, as much as 80 percent of the profit is generated by serving 20 percent of the customers, while 80 percent of the revenue is generated by 20 percent of the products.

Apparently equity market returns also show a Pareto distribution.

source: Longboard Funds

Most workers make tradeoffs between housing costs and commuting time and hassle that show a Pareto distribution.

In  the healthcare domain, the sickest 20 percent generate 79 percent of health care costs.

In the domain of athletic training roughly 20 percent of the exercises and habits have 80 percent of the impact

It is likely that similar Pareto distributions exist for all forms of internet access and communications infrastructure, where 80 percent of the value comes from 20 percent of the decisions or instances.

As a practical matter, Pareto means I have to spend more of my research time on the relatively few connectivity firms that generate 80 percent of the revenue in any particular market, even if much of my work has included managed services providers, distribution partners and others that are part of the 80 percent of firms that generate 20 percent of the revenue.

Android 9 Pie Update Incorporates Machine Learning (Artificial Intelligence)

The latest update of the Android operating system, now rolling out, incorporates a fair amount of machine learning to predict what users will want to do next (less scrolling). 

Android 9 also uses machine learning to adapt battery use to personal usage patterns (brightness levels, for example). 

Mobile Video Could Boost ARPU 14% to 24%

In the U.S. market, the Federal Communications Commission says the average revenue per mobile user was about $35 per month in 2017, implying annual revenue of about $420 per user.

So an incremental $11 or $12 per month, per account, is a significant opportunity, boosting per-user revenue perhaps 24 percent to 26 percent. At 50-percent adoption rates, at a monthly rate of $11, ARPU could climb to $40.50 a month, if a particular mobile operator can sell a subscription service to half its installed base.

source: FCC

New Multi-Billion Dollar Mobile Markets are Few, Far Between

There are not many multi-billion-dollar new markets mobile operators can easily create.

Verizon’s media operations generate perhaps $8 billion annually, by way of comparison, based on revenues from The Huffington Post, TechCrunch and Yahoo Sports and other assets acquired for about $9 billion as part of the AOL and Yahoo buys.

And that is generally considered to be a failure.

Consider that the whole U.S. software-defined wide area network market might be worth $2 billion annually. And SD-WAN arguably is the most-strategic product in the WAN provider business.

The U.S. “unified communications” business, including on-premises switches, conferencing services, hosted services, software and trunking services, might be in the $8 billion to $19 billion range per year, depending on which estimate, with which constituent elements, is used.

source: Grand View Research

The point is that it is rather hard to create a new revenue stream measuring in the billions of dollars, in the U.S. market.

Linear video subscriptions now are worth $91 billion annually in the U.S. market.  

source: FCC

And that, quite simply, is why many mobile operators see subscription video as a meaningful new revenue source. It is a huge generator of revenue on a scale that few other products can match.

Mobile Networks Cost 9% of Revenue, Cable 45% of Revenue, Telco 40% of Revenue in Cana

It would be hard to point to any single graph that better shows the relative cost of mobile and fixed (cabled) networks, measured in terms of the percent of revenue expended on each type of network.

With the caveat that the mobile figures might be periodically higher when a next-generation network is under construction, ongoing capital investment by Canadian mobile operators is about 10 percent of revenues.

Canadian telcos tend to invest about 35 percent to 40 percent of revenues on their networks, while cable operators tend to invest about 45 percent of revenue, according to a report by Canadian regulator CRTC. Those figures probably require some explanation.

Cable networks have historically been less capital intensive than telecom networks. But cable operators also are energetically investing in gigabit internet access, while most telcos are more measured in their spending. In part, that is because of the high cost of upgrading copper to fiber access facilities.

But telcos likely also are measured in their assessment of revenue upside from fiber upgrades. In other words, they might rationally conclude that there is no business case for rapid fiber upgrades, especially given revenue declines and a cable TV advantage in internet access and video services.

source: CRTC

Mobile services revenue generated 51 percent of all service provider revenue in Canada in 2017, the CRTC reports. Fixed network internet access generated 23 percent of total revenue.

Only mobility services and fixed network internet access sources grew; all others declined in 2017, CRTC says. And 34 percent of total revenue is earned by cable TV companies.

source: CRTC

Abstraction (SDN, NFV) Might Enable New Federated Mobile Networks

“Abstraction” is what makes any over-the-top internet application or service possible and viable. So the possible coming issue is whether abstraction of network functions flowing from network functions virtualization and software defined networks might lead to the feasibility of “abstracted” or “virtual” service providers.

One example of the business value of abstraction is the ability to construct a global business, when scale matters. Compare the operations of webscale firms such as Amazon, Facebook or Google with the operations of mobile service providers, for example. By 2014, it was clear that the app-based providers had gotten more scale than nearly every mobile service provider.

source: Nokia Bell Labs

Think about a mobile virtual network operator model “on steroids,” where a service provider can assemble a full network and services from one or more underlying networks, as Google Fi, which dynamically connects its mobile users using Wi-Fi, Sprint or T-Mobile US networks.

Future networks, built on NFV and SDN principles, could well enable new service providers who federate resources from many physical networks.

source: Nokia Bell Labs

It is the AWS model applied to the connectivity business. As enterprises switched from reliance on owned infrastructure to using cloud services platforms, so it might eventually make financial sense for some connectivity providers to create their own networks on a virtual basis, especially when aggregating assets from multiple physical networks has value.

In other words, it might be possible for entities such as Amazon and others to operate their own access networks, the same way they now operate their own wide area networks. Google actually already operates its own mobile and fixed networks.

source: Nokia Bell Labs

In 2019, North America Will Represent 47% of Global 5G Subscriptions

5G Americas predicts that there will be 336,000 5G connections in North America by the end of 2019, representing 47 percent of global 5G connections. Longer term, the Ovum numbers predict 186 million 5G connections in North America by 2023, accounting for 32 percent of the expected 1.3 billion connections worldwide.

To put that in perspective, the United States has about five percent of the world’s mobile accounts of about 7.8 billion mobile subscriptions.  

source: Ovum

"Fake 5G" and Real Demand

Most discussions of bandwidth for consumer apps are a matter of overkill. Observers complain about “fake 5G,” “slow internet speeds” and “high prices” for internet access. There are some geographic areas (mostly rural, but including some parts of cities, apparently) where absolute speed is an issue.

But most of the time, speed does not prevent people from using all the apps they want, with acceptable user experience. The reason is simply that most apps require very little bandwidth, and that the key variable is simply the total number of simultaneous users or devices.

Generally speaking, 20 Mbps per user or device is quite sufficient.

Number of devices	Use Cases	Recommended Download Speed
1-2	Web surfing, email, social networking, moderate video	Up to 25 Mbps
3-5	Online multiplayer gaming, 4K streaming	50 - 100 Mbps
More than 5	All of the above plus sharing large files and live streaming video.	150 to 200 Mbps

source: Toms Guide

The evidence that most users require relatively modest bandwidth (20 Mbps per user, for example), comes in part from actual consumer buying behavior. As a recent survey of small rural telcos shows, consumers do not generally buy the fastest-available tier of service.

Even when gigabit connections are available, few consumers buy them. Among rural telcos offering internet access, some 23 percent of all connections made available by these rural service providers offer at least a 1,000-Mbps connection. Yet only two percent of actual subscriptions are for gigabit services.

Another 34 percent of connections offer speeds from 100 Mbps to about 999 Mbps. Yet less than 14 percent of customers buy such tiers of service.

Source: NTCA data , IP Carrier analysis  

Some will lament the fact that even when gigabit (or any other very-fast) internet access service is available, most consumers do not seem to buy them, when there also are choices of other services that cost less, but are not as fast. In other words, there typically is some gap between the availability of a service and consumer willingness to buy.

One logical conclusion is that people find the services they buy suit their needs, and that many consumers see relatively little additional advantage to buying a gigabit or other very-fast tier of service.

In the United Kingdom, for example, 94 percent of U.K. homes and businesses are in areas where fixed network broadband operating at 30 Mbps or faster is available, according to Ofcom. In such areas, just 45 percent of homes buy a service operating at 30 Mbps or faster.

In other words, only half of customers will buy services of 30 Mbps. And yet in other markets, such as the United States, where some observers decry the “slow speeds” and “high cost” of internet access, the same results are seen. Most consumers choose not to buy gigabit services.

And such data underscores an “iron law” of the consumer connectivity business. People are only going to spend so much for connectivity services. And in many markets, we have long passed the point where actual speed was an impediment to user experience.

5G Fixed Wireless is More Important Than Many Believe

With the commercial launch of 5G services, commentary often runs to an extreme. Some hype the longer-term new use cases, while others say not much will change, for the typical consumer.

Both viewpoints have some merit. Longer term, 5G promises brand-new use cases, largely around internet of things. Near term, there are not many 5G phones, few places to use 5G networks and few--if any--use cases that actually can avail themselves of 5G network capabilities.

source: Deloitte

Fixed wireless, perhaps oddly, is among the few new use cases 5G networks will enable. Mobile internet access and personal hotspots will be among the use cases for early 5G users, but those are existing use cases.

Deloitte Global predicts that 20 percent of North Americans with internet access will get all of their home data access from mobile networks in 2018. The percentage of mobile-only customers might grow to as much as 30 percent to 40 percent of the population by 2022, Deloitte Global estimates, including both mobile-only and mobile operator fixed wireless users.

source: Deloitte

For some internet service providers, 5G fixed wireless offers an attractive new revenue source, allowing market share capture from fixed network service suppliers. That strategy worked well in many other parts of the communications business.

MCI took long distance calling share from AT&T; satellite video providers took share from cable TV companies; cable companies took voice share from telcos; telcos took video share from cable companies; mobile firms took voice share from fixed network providers.

Some mobile ISPs will be able to take advantage of the 5G fixed wireless opportunity; others largely will not.

5G Might Not be Something Most Consumers Buy in 2019

“There will be many commercial 5G launches but consumers will not notice,” Analysys Mason argues. They probably are mostly right about 2019 and likely 2020 as well. It is not simply the relative paucity of 5G phones--including the lack of Apple iPhone support--but other developments in 4G which make LTE more attractive, with a more-robust supply of devices and therefore choice.

How Much Difference in 5G User Experience?

The difference in experience between advanced forms of 4G and 5G networks, with the possible exceptions of big downloads and experience compared to public Wi-Fi, might not be so visible to most users, most of the time. There will be significant speed advantages, to be sure.  

Compared to standard 4G, there will be significant speed differences for browsing, some simulations by Qualcomm suggest.

Browsing download speeds increasing from 56 Mbps for the median 4G user to more than 490 Mbps for the median 5G user, a gain of approximately 900 percent, Qualcomm says.

The good news:if download speed is a constraint, it will not be, in the 5G context.

The bad news: beyond a fairly-low limit--perhaps 10 Mbps to 20 Mbps--any single user will not experience noticeable better experience, on either latency or bandwidth dimensions.

source: O'Reilly Media

Latency also decreased about seven times, with median browsing download latency reduced from 116ms to 17ms.

A study by Signals Research Group, intended to show the benefits of gigabit 4G, suggests that gigabit 4G, similar in many ways to 5G, will tend to show benefits for downloading of big files, especially compared to public Wi-Fi.

Signals Research Group

The issue is that once access speed (the 4G or 5G connection) is improved, experience issue elsewhere in the value chain will be highlighted. Even if the local connection on both ends of any server interaction are very fast, eliminated latency or bandwidth as constraints on experience, server latency on the far end will still exist, and will become the likely experience bottleneck.

To be sure, mobile device latency  lags that of fixed networks, generally speaking. But 4G was better than 3G, and 5G will be substantially better than 4G on that score.

India Moves to Eliminate Interconnection Fees

India’s TRAI has been working to lower the mobile phone interconnection usage charge since 2016, ordering a reduction in such interconnection fees in 2017, and planning to eliminate the charges altogether by 2020.

Upon such minutiae do business models turn, which is why some mobile operators oppose reducing the interconnection charge to zero. Though in principle, when traffic is symmetrical, such charges result in a net zero revenue impact, termination charges produce revenue.

So some argue that the elimination of interconnection charges should not set the IUC at zero, which is what systems known as “bill and keep” produce.

In principle, lower or zero levels for such fees reduce the cost floor for making calls and using networks.

Bill and keep is a pricing arrangement for the interconnection of two telecommunications networks under which the reciprocal call termination charge is zero. That is a change from the older practice of terminating charges paid by the originating network to the terminating network.

That call termination charges exist at all is a reflection of the notion that it costs a non-zero amount of money to run a network, and that a network completing a call should be paid by the originating network.

There are business implications when traffic is symmetrical, since all major parties basically experience the same amount of terminating calls, but mostly of the “higher cost to consumer” sort. When termination charges are paid, that cost is reflected in the retail rates for using a network.

On October 27, 2011, the U.S. Federal Communications Commission adopted a bill-and-keep framework for all telecommunications traffic, in large part because of price arbitrage and practices such as traffic pumping and phantom traffic.

Traffic pumping, also known as access stimulation, is a practice by which some local exchange telephone companies in rural areas of the United States inflated the volume of incoming calls to their networks, profiting from intercarrier compensation fees.

Phantom traffic includes calls that do not have the identification information used by carriers to levy interconnection fees. That obviously meant that some firms profited by avoiding termination fees.  

It is arcane, but rules on interconnection do directly affect service provider business models.