Saturday, March 31, 2018

5G is Like the Tail on a Dog

5G is to networking as telecom is to the internet. That is to say, 5G is part of a larger shift of networking as "telecom" has become a tail on the internet dog.

Specifically, 5G is part of a larger transformation of global public networks involving much lower latency, much more virtualization and new roles for data centers. Those features, in turn, are required to lower the cost of running networks as well as create the foundation for new categories of services that will drive incremental revenue at scale.

Extremely low latency, high connection density, high reliability and gigabit speeds are driving the design of the whole new architecture, with implications for access, cloud computing and virtualization.

Virtualization is a key change, with separation of control and signaling functions from delivery of end user traffic becoming key. Lower cost is among the expected outcomes. Greater flexibility also is anticipated.

What might be relatively unexpected is that virtualization will increase the ability to create new virtual networks on a wider scale. That could have key implications for new suppliers, including suppliers that only want to create large virtual WANs to support internal
business requirements.

The analogy here is the shift of wide area networking from “mostly public” to “mostly private.” In other words, on many routes, enterprises carry most of the traffic, but only to support their own internal operations. Major app providers, in other words, have vertically integrated WAN services.

That has business model implications both for the enterprises and suppliers of WAN services.

Also, to support new ultra-low-latency apps, data center functions will emerge at the edge of the network. In that sense, 5G and edge computing are parts of a larger trend, the creation of next-generation networks that are both virtualized and built around latency performance.

That stringent latency performance will be needed to support emerging new applications in the connected car, internet of things and real-time application areas. Bandwidth will matter, but latency arguably is the key new capability.   

Friday, March 30, 2018

The Prejudice Against Bigness

It would be reasonable enough to argue that there is an almost-instinctive distrust of "bigness" in business, as such bigness is presumed to be responsible for destroying the fortunes of small and local businesses.

It must also be said that it is consumers who propel the rise of "big businesses," since it is consumers who prefer the products, prices or other attributes of bigger businesses, over those of smaller suppliers.

We might be tempted to decry bigness, but consumer choice is what produces bigness. And that means bigness is not always and inevitably bad.

Sometimes, in fact, bigness might be a survival requirement. Consider some foundations of business in the internet era. Among the key trends are a few that define internet market dynamics:

* Pricing and margin pressure
* New competitors from outside the existing value chain
* Winner take all market share dynamics
* Disintermediation of distributors in the value chain

In case you somehow missed the trend, the internet leads to lower retail prices and lower supplier profits, as price transparency increases and distributors are removed from the value chain. 

Faced with lower retail prices, suppliers can respond in a few ways. They can sell more units at the lower prices, which requires more scale. Firms can sell different products, which means moving into new or different parts of the value chain. And firms can take on additional roles within value chains, increasing the ways they make revenue. 

The point is that all that tends to imply firms get bigger. 

Also, the internet allows firms outside the existing industries, possibly operating with different value and revenue models, to enter existing businesses and disrupt them. That creates additional pressure on sales volume, prices and profits. 

Market disruptors also frequently rely on internet capabilities to remove cost from the value chain by eliminating steps and roles in the distribution process, also after changing costs in the production part of the value chain. 

The "winner take all" nature of many internet-reliant businesses and industries also contributes to bigness. 

So we need to be careful not to confuse bigness, which might well be a requirement for firm or industry survival, with actual anti-competitive behavior on the part of big firms. Abuse always is a possibility when power exists in a market. But abuse is not automatic, and hopefully not a routine outcome.

Nor should we automatically penalize firms for delivering higher quality products at lower prices. That is the outcome we are supposed to gain from competition. 

But sometimes competition can only be fostered by big firms. Break up the big firms and you also possibly destroy ability to innovate and create even more consumer benefits. 

Bigness is not inherently bad. 


Antitrust The Wrong Solution for the Wrong Problem

At the risk of oversimplifying, the apparently-growing sense that “something has to be done” about the size of today’s firms (financial, retail, telecom, internet apps and so forth) is likely ill considered.

As profit is wrung out of all value chains affected by the internet, firm revenue and profits fall, if not to zero, then always in that direction. There is only a few long-term solutions for such margin compression: additional scale in existing businesses, and a move into new businesses, elsewhere in the value chain.

Both strategies require that firms get bigger. So attacking "bigness" also means attacking chances for firm survival.

The demand for scale is--virtually all agree--a byproduct and necessity in an era of price transparency, lower protections from market entry by “outsiders,” falling prices and profits in virtually all incumbent businesses and markets.

We cannot easily repeal the economic impact of the internet, even if we wanted to do so. And make no mistake, the internet will reshape nearly every industry in certain ways, the cumulative effect of which is to wipe out profit margins. That, in turn, drives the need for scale.

Also, economic eras change. And that is why, in retrospect, lots of antitrust action seems not to work. It ultimately did not matter what we did about the size of the oil or steel or auto industries, as they were destined to lose their place as the economic engines anyhow.

The big revision of U.S. telecommunications law (the Telecom Act of 1996) aimed to introduce competition for voice services, precisely at the point that voice was shifting to mobile delivery and the rest of telecom was shifting to the internet.

Some of us question the long term value of antitrust action applied to Microsoft (the internet era displaced the PC era anyhow).

And, for such reasons, some of us are quite skeptical about the value of antitrust action against AT&T, Amazon, Facebook, Google or Apple, especially when the types of action we see are instances of vertical integration, which does not lessen competition in existing markets.

The big problem is that we humans are always “fighting the last war,” seeing dangers that already are destined to pass. Consider the U.S. Department of Justice blocking of two big mergers in the health insurance industry, and the blocking of mergers in the retail pharmacy industry, because of concerns about excessive concentration of power that seem, in retrospect, unnecessary.

Walgreens abandoned its effort to buy Rite Aid in 2017, after the Federal Trade Commission said it would review the proposed transaction on competitive grounds.

That was a horizontal merger. Now, however, a big possible wave of vertical transactions might happen, as the pharmacy benefits and insurance industries face the new threat of Amazon entry into the ecosystem.

That already is raising calls for antitrust action to block vertical deals such as insurer Cigna merging with Express Scripts. Pharmacy CVS is merging with insurer Aetna.

Other deals might happen as well, including Walmart and Humana.

The point is that, in the internet era, scale becomes a necessity, to deal with lower revenues and profits, loss of value and business model disruption.

The point is that firms and industries need to change in the internet era, and gaining scale, plus entering new parts of the ecosystem, are often necessary survival and relevance actions.

We may generally believe “bigness is bad,” but bigness might well be necessary when profits, margins, value and roles are compressed.


What is “big” in a legacy context might well be “small” in a broader internet context, where the definition of “market we are in” changes, or has to change.

Some might say antitrust concerns considered “reasonable” in a horizontal context actually make no sense in a “vertically changing” industry or market. That might especially be true when whole markets are shrinking, disappearing or are threatened with that eventuality, as value shifts.

That is why some of us view the U.S. Department of Justice antitrust case against the vertical AT&T merger with Time Warner unfortunate and misplaced. The industry is going to have to change, vertically, to survive, as AT&T’s major competitors already are doing.

Vertical integration already is happening in many parts of the internet ecosystem, meaning the question who are my competitors changes.   



Thursday, March 29, 2018

5G Will Accelerate Mobile and Wireless Substitution for Internet Access

Up to this point, 4G has been a primary method of internet access for a substantial, but still minority of usage in developed markets, though it arguably has been a primary or exclusive form of access in developing markets.

But the 5G era is likely to accelerate those trends in a major way. For the first time, mobile or 5G-based fixed wireless networks will offer speeds and retail prices as good--or better--than fixed alternatives, with latency performance that is better .

Many of us would not be at all surprised if wireless substitution as much as doubled over the first decade of 5G commercial service. In some developed markets, that could mean that wireless access takes as 20 percent share of the residential internet access market. In other markets, wireless internet access could reach 30 percent or higher.


You might be tempted to think that it mostly is lower-income people who use their mobile devices as the sole means of internet access. In many countries, substantial numbers of people--rich and poor--do so.

While it is true that, in the United Kingdom, relatively few people actually are “mobile only for internet access (2.5 percent of the richest consumer households; perhaps seven percent of the poorest  households), in Canada, 20 percent of the richest households are mobile only for internet access, while about 30 percent of the poorest households are mobile only.

Households in rural areas also are more likely to rely on mobile networks for their internet access requirements as well.

Single-person households and households with younger consumers also are more likely to be mobile only for internet access.

Less than a tenth of people in France and the UK were mobile-only, but in Turkey the figure was more than three times higher.

In Latin America, for example, Deloitte Brazil believes that over a third of all homes in Brazil were mobile data only. And in China, a fifth of the online user base (rather than households) were mobile-only as of 2016.

In Tokyo, where fiber optic connections are widely available, hundreds of thousands of homes (or about five percent) are relying on only mobile in 2017.

In Latin America, Deloitte Brazil believes that over a third of all homes in Brazil were mobile data only.

Better mobile networks, especially 4G LTE, have made mobile internet user experience more like fixed access.

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.

Deloitte Global further predicts that a mixture of mobile and fixed wireless access technologies could lead to 30 percent to 40 percent of the population relying on wireless for data at home by 2022, an increase from 10 percent in 2013.


What is unclear is how wireless access might change over the next five years. Deloitte argues that, by 2022, wireless home internet solutions will grow both at the low end of the market (homes using relatively little data) and portions of the market that otherwise might have purchased a fixed connection.


Full Duplex for Cable TV is Another Example of Virtualization

Virtualization is becoming a core feature of next-generation networks. Consider how the cable TV industry is looking to use a form of virtualization in its access networks. Traditionally, cable TV networks have used frequency division for both upstream and downstream communications.

Telco networks have used time division. But the latest development is that cable TV networks are looking at a way to use “full duplex” techniques in place of frequency division, a capability that could reduce the cost of rebuilding physical networks to boost bandwidth.

To use full duplex, a hybrid fiber coax network would have to operate in “passive HFC” mode, which means no use of active signal repeaters in the distribution network, in all likelihood.

It is hard to see how full duplex could work through any set of bandpass filters, standard in a radio frequency amplifier.

Full duplex would allow all spectrum to be used for upstream or downstream communications at the same time. And that would be a form of virtualization, separating logical channels from underlying physical network frequencies and channels.  

Today’s Technology Duplex: Time Division Duplex
Frequency_Division_Duplex
Today’s Technology Duplex: Frequency Division Duplex
Time_Division_Duplex
Full_Duplex_DOCSIS


How Does Private LTE (CBRS) Affect the WAN Business?

It is by no means clear how the availability of private (enterprise) 4G Long Term Evolution will affect the broader mobile market.

The ability to create indoor or campus LTE networks essentially complicates some elements of enterprise premises communications strategy. Where the typical network has been a cabled local area network, now supplemented with Wi-Fi and mobile access, the future might also include Wi-Gig as well.


On one hand, enterprise indoor 4G, based on use of small cells, could help provide better mobile coverage indoors, with less capital investment than might otherwise be required. In other cases there might be some incremental increase in operating costs, especially when mobile operator access to the indoor networks is a lease arrangement with the private LTE operator.

On the other hand, the indoor LTE network could well, over time, build beyond support for mobile phones and include internet of things sensors. That could affect mobile business revenues in a variety of ways, from displacement to sharing to augmentation.

The private LTE operator could be a buyer of access and transport; a competitor for IoT services; a partner for providing better indoor coverage or platforms, possibly all at once.


In other words, private LTE might simply be another form of indoor local area network that aggregates local traffic for transmission over wide area networks, as local area networks (Wi-Fi, for example) do. In this scenario, private LTE is simply a way of supporting employee, partner and customer use of existing services and networks.


To some extent, especially in smaller venues, private LTE might be a service sold to building managers or tenants, much as airport or hotel Wi-Fi service is a business for some suppliers.

In all cases, private LTE could establish a new role for LTE networks in the role of local area network, as Wi-Fi now is used. If you think about how Wi-Fi affects the WAN service providers, there are stimulative effects, creating additional demand for internet access services.

Private LTE should also help solve indoor signal strength issues, as has been true for mobile internet access functions inside buildings.

If you think about private LTE networks supported by new spectrum such as Citizens Broadband Radio Service as a new form of local area network platform, akin to Wi-Fi, you can figure out how it might be used, and what the implications might be.

But it always is fair to remember the fundamental distinction between local and wide area networks. WAN services are fundamental for service provider business models. LANs such as Wi-Fi have become more important functional parts of the WAN business, though, allowing service providers to provide better service (data offload and signal strength).

Private LTE will probably have much the same effect, even as some incremental new business opportunities could be created for suppliers of indoor communications services.

Wednesday, March 28, 2018

What is AI Job Impact? What's the Best Balance?

One possibly hopeful predication about how automation (artificial intelligence and machine learning) and human jobs can coexist is that there will some job loss, some job redefinition and some job creation.

In fact, some might argue that an optimal balance is to apply artificial intelligence and machine learning to some degree, but not too much.

To be sure, some job categories will be more at risk, according to some researchers. As a rule, some believe, jobs most at risk to automation are those which are routine.

It probably also is fair to note that automation tends to redefine jobs as well, so overall impact is complex.

 






U.S. Will Adopt 5G Very Fast, Says GSMA

It is not hard to find skepticism about the size of 5G markets, for good reason. Not every market is equally well disposed to generate new revenue sources, at scale, and not every set of service providers in every market has the same level of actual need to find those new revenue sources.

That is why virtually everyone expects 5G to become a commercial reality first in just a handful of countries: the United States, Japan, European Union countries and China. Developed nations plus China, in other words. And some might question how fast most EU markets will adopt.


In part, that early deployment pattern is driven by expectations of financial upside from new applications not possible with 4G. GSMA Intelligence also notes some particularities of the U.S. market that also create a fertile environment.

In other words, opportunities might exist to drive aggressive 5G adoption in the U.S. market.

GSMA Intelligence notes that U.S. customers have been robust adopters of “digital” services using their mobile devices, ranging from internet-based messaging and social media to entertainment content, as well as using e-commerce.

That’s the “pull” of 5G. There also is the “push.” U.S. mobile service revenues have been steadily declining since 2010, and might go negative in 2019. That means the search for additional new revenues is imperative.

So all four leading U.S. mobile operators will be offering 5G mobile services in 2019, with AT&T and Verizon launching mobile 5G services in selected markets in 2018.

According to GSMA Intelligence forecasts, the United States will experience one of the fastest customer migrations to 5G in the world, with 5G reaching 100 million mobile connections in early 2023.

That will make 5G the leading mobile network technology in the United States by 2025, with more than 190 million 5G connections (accounting for around half of total mobile connections), excluding use of 5G to support fixed operations.

But more than supply will drive that deployment speed.

Aggressive millimeter wave spectrum allocations will be key enablers, especially for AT&T and Verizon.

But there are other fundamental reasons for the push into 5G. Simply put, the U.S. mobile market has exhausted its growth model, based on 4G. If one extrapolates from steadily declining revenue growth since 2010, it is possible revenue growth in the core mobile business might actually go negative in 2019.

Since revenue growth is a virtual necessity for firms that are publicly owned, that represents a crisis.


Though new revenue sources will have to be created, 5G offers some opportunity for services that actually have not existed on earlier platforms, including 4G. Though everyone expects early revenue upside to come in the area of “enhanced” (much faster) internet access, that expectation must be qualified.

Since 4G adoption is virtually ubiquitous in the U.S. mobile business, customers who adopt 5G access will, by definition, be replacing primary reliance on 4G. So there is substitution going on, replacing 4G accounts with 5G.

There may or may not be incremental upside from such substitution. In the early days, some incremental upside is likely.

Longer term, the hope is that brand new revenue streams can be created in the internet of things (sensor communications) area. There also is hope some new classes of apps can be created in the ultra-low-latency area. Connected cars or 4K and 8K video, plus other latency-dependent apps such as remote surgery come to mind.

The important point is that U.S. mobile operators are going to be aggressive about 5G because the 4G business model has become exhausted.

Internet Access Universal Service is Always an Issue for the Last 2% of Locations

Universal funding of “essential” telecom services always is difficult and expensive, since, by definition, there often is no private sector business model for rural areas. In large part, that is because there are too few potential customers, in relation to the cost of assets to serve those potential customers. In the continent-sized U.S. market, for example, the universal service problem largely is a matter of rural areas.  

Assuming a standard fixed network investment cost, that might not produce a positive business case over a 20-year period, the U.S. Federal Communications Commission has suggested.  And almost nobody makes investments with a 20-year payback in telecom, anymore. That is tantamount to “no investment return.”

In the United States, the cost of serving the last one percent of locations is astronomical, for example.


High infrastructure costs  are among the reasons wireless access is likely to play a bigger role in such universal service plans. Fixed networks cost too much in rural areas, in other words.

Consider many U.S. states where rural population density ranges between 50 and 60 locations per square mile, and ignore the vast western regions east of the Pacific coast range and west of the Mississippi River or 100th meridian,  where population density can easily range in the low single digits per square mile.

Assume 55 locations per square mile, and two fixed network suppliers in each area. That means a theoretical maximum of 27 customers per square mile, if buying is at 100 percent. Assume for the moment that buying rates really are at 100 percent. Two equally skilled competitors might expect to split the market, so each provider, theoretically, gets 27 accounts per square mile.

At average revenue of perhaps $75 a month (perhaps a generous assumption), that means total revenue of about $2025 a month, per square mile, or $24,300 per year, for all the customers in a square mile.

The network reaching all homes in that square mile might cost an average of $23,500 per home, or about $1.3 million.

At 50 percent adoption, that works out to network costs of roughly $47,000 per account in a square mile, against revenue of $900 per account, per year. Over 10 years, revenue per account amounts to $9,000. Over 20 years, revenue might be a bit more than $18,000.

The business case does not exist, without subsidies.

In the United Kingdom, Ofcom, the U.K. communications regulator, is planning a minimum 10 Mbps internet access floor for all locations in the United Kingdom. That probably is an issue for the last one to two percent or so of U.K. locations.

Ofcom estimates a per-location cost of £3,400, enabling coverage to around 99.8 percent of premises. Consumers outside this threshold will be able to get a satellite connection. That figure is a blend of higher-cost (and relatively speaking) lower-cost locations.

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...