4G Fixed Wireless Closes Cost Gap with Cabled Networks

One traditional reason mobile networks have not been direct substitutes for fixed (cabled) networks for internet access is the value-price relationship. Where mobile network data consumption has been priced at about $9 to $10 per Gbyte, fixed networks have offered gigabytes at less than $1 per gigabyte (tariffs sometimes below 20 cents per Gbyte).

Basically, mobile data has cost 20 times (actual usage) to 60 times (retail price compared to usage limit) that of cabled access. Leaving speed out of the analysis, mobile data has not been affordable, compared to equivalent amounts of cabled network usage (between 160 Gbytes to 200 Gbytes, for example).

Speed also has been an issue, as mobile networks traditionally have been slower than fixed (cabled) networks.

But AT&T already is moving to eliminate much of the “cost of data” gap, using its 4G network, in rural areas, in fixed mode.

The 10 Mbps AT&T fixed wireless service includes a usage allowance of 160 Gbytes per month, and actually uses the Long Term Evolution (4G ) network in a fixed mode.  For most consumers, that will close the “cost per Gigabyte” problem.

Speed remains an issue, however, at least until the 5G network and millimeter wave bandwidth is made available. Still, the AT&T deployments show that mobile networks can move closer to comparability with cabled networks, using 4G, and could close the gap entirely in the 5G era.

AT&T has launched its fixed wireless service in Georgia, planning to reach 26,000 locations by end of 2017 and over 67,000 locations across Georgia by 2020.

AT&T plans to reach over 400,000 locations nationwide using fixed wireless by the end of 2017, and over 1.1 million locations nationwide by 2020.

Google and Facebook Will Control QoS on Network Edge

“As we know, Google’s strategy is increasingly to own both ends of the communications pipe,” Openwave Mobility says, noting the use by Google of new transport protocols that inherently include encryption and latency control. In fact, says Openwave, anecdotal reports suggest end users on Google Quick UDP Internet Connections already experience 30 percent less video buffering than other users nor viewing videos using QUIC.

That is vitally important for several reasons. First, Google and Facebook increasingly are in charge of quality of service mechanisms of their own traffic. Also, any so-called “internet fast lanes” will be impossible for internet service providers to create, or of little to no value to Google or Facebook.

For starters, Facebook and Google will control QoS themselves. Secondly, ISPs cannot “speed up” packets when they cannot determine the owner of a packet and its media type, since the packets are encrypted.

So with the anticipated and breathless worries about the coming of “internet fast lanes” if common carrier regulation of internet access services are  removed, it might be helpful to remember that the technical ability to create quality of service tiers of service (the dreaded internet fast lanes that reduce latency) extending all the way to end users might not actually exist, in a majority of cases.

The reason is encryption. To do anything on a selective basis to a packet, an ISP essentially needs the ability to identify the owner of a packet and the media type of a packet.

With so much traffic already encrypted, that is impossible at least 70 percent of the time, already. According to Openwave Mobility, about 75 percent of traffic already is encrypted.  

By perhaps 2018, 90 percent or more of all packets will be encrypted, Openwave predicts.

Google QUIC and Facebook Zero Protocol (0-RTT), for example,  represent 27 percent of the traffic in mobile data networks already.

source: Openwave Mobility

Additionally, new protocols such as Quick UDP Internet Connections (QUIC)  intrinsically include security protection equivalent to TLS/SSL, along with reduced connection, transport latency and bandwidth estimation in each direction to avoid congestion.

QUIC also provides mechanisms for congestion avoidance algorithms, putting control into application space at both endpoints. In other words, the app provider can supply its own congestion control, and does not need to rely on a transport or access provider to do so, to obtain the benefits of congestion control.

source: Openwave Mobility

The ability to create “internet fast lanes” requires visiibility at the packet level, a condition that largely does not exist. When the consumer internet already has “gone dark,” such visibility is not possible.

Therefore, no ability to create a new consumer quality of service tier actually exists, no matter what many argue is the new danger if common carrier regulation of internet access is dropped.

Mobile Revenue Per Bit is a Big Problem, Getting Bigger

Revenue per bit is a big problem for access providers (internet service providers)  today, and will be a bigger problem tomorrow, especially on mobile networks.

Video has the absolute lowest revenue per bit, and therefore potential profit per bit, of any traffic type. Video also dominates data volume on global networks, but has revenue per bit perhaps two orders of magnitude (100 times) less than voice, for example.

Those reductions in price per unit sold would be nearly catastrophic in any business, and tell you most of what you would need to know about the direction of network capital and operating costs to achieve sustainability.

At the same time, those catastrophically-lower revenue per bit measures also tell you why surviving tier-one service providers must (not “should”) find big, new revenue sources.

Text messaging has in the past had the highest revenue per bit, followed by voice services. More recently, as text messaging prices have collapsed, voice probably has the highest revenue per bit.

Video always has had low revenue per bit, in large part because, as a media type, it requires so much bandwidth, while revenue is capped by consumer willingness to pay. Assume the average TV viewer has the screen turned on for five hours a day.

That works out to 150 hours a month. Assume an hour of streaming (or broadcasting, in the analog world) consumes about one gigabyte per hour. That represents, for one person, consumption of perhaps 150 Gbytes. Assume overall household consumption of 200 Gbytes, and a monthly data cost of $50 per month.

That suggests a cost--to watch 150 hours of video--of about 33 cents per gigabyte. Assume a mobile or fixed line account represents about 350 minutes. Assume the monthly recurring cost of having voice features on a mobile phone is about $20.

Assume data consumption for 350 minutes (5.8 hours a month) is about 21 Mbytes per hour, or roughly 122 Mbytes per month. That implies a cost of roughly $164 per gigabyte.

In other words, video revenue for the access provider is perhaps 33 cents per gigabyte, while voice generates perhaps $164 per gigabyte. These days, unlimited domestic texting is often a feature, not a revenue driver, generating zero revenue per gigabyte.

Since video now is the application that now drives global network traffic, you can understand the need to reduce cost per bit in the network, and in all network-related operations, to sustain the networks, if revenue per bit is declining.  

source: Ericsson

Will Only 10% of 2023 IoT Connections Use Mobile Network?

As important as the internet of things is likely to become, as a driver of incremental new 5G network revenue, it remains unclear just how big a revenue boost the “access” connections will be. According to Ericsson, some 20 billion connected IoT devices will be use by 2023.

But only 1.8 billion of those IoT devices will use mobile connections in 2023.

In other words, less than 10 percent of total IoT sensor connections will use the mobile network. That explains the interest in specialized low-power WANs.

Most of the IoT devices (perhaps 90 percent)  will connect using other local and short-range networks.

source: Ericsson

5G Evolutionary or Revolutionary? Yes.

Every important new technology gets hyped, for good reasons. Suppliers see big markets, and have every reason to promote the upside, to boost sales and grab leadership of the expected big new markets.

Futurists and journalists have nothing to talk about if they cannot speculate on the transforming potential of general purpose (GPTs) and other technologies. But GPTs exist. They transform work, life and economies.

Also, no matter how hard we try, humans are notably poor at making accurate forecasts accurate forecasts about the future. We rarely do much better making more mundane estimates of future prices, who your competitors will be or what products you will be selling in a decade.    

Beyond that, important new technology follows a pattern described by the S curve, which virtually guarantees that any systemically important new technology will take some time to develop. That very process also ensures, however, that impact is virtually always less than expected at first, and far more important, later.

The point is that it is almost pointless to debate whether some new technology actually will become a general purpose technology that transforms life and work: we often cannot know that, at first.

And even if a GPT is coming, its impact will not happen in the first years, but only over decades.

For that reason, even debates about whether the impact is evolutionary or revolutionary are nearly pointless. Even if a GPT is coming, its initial impact will seem quite evolutionary, and often quite limited. Only decades later will the larger impact be so obvious.

With regard to 5G, all these caveats are relevant. There has been significant uncertainty about changes in revenue models and lead apps for every next generation mobile network. The immediate shifts were always evolutionary.

Expected innovations in consumer behavior and apps always took some time to develop, and perhaps the most-important shifts occurred over several different network generations. Some potentially-important “new data apps” we might have conceived of in the 3G era might actually emerge only in the 5G era.

And, to be sure, most of the actually realized initial revenue from 5G will be 4G usage that shifts to 5G. So yes, 5G will appear to be quite evolutionary. In some sense, that always is the case for next generation mobile networks.

But there are some potential new twists. If, as many expect, the global telecom business radically consolidates, if 4G improves enough to provide 5G type features (bandwidth and latency), if the new internet of things revenue upside accrues mostly to a smallish number of big tier-one operators, then it would be rational for executives to set lower expectations about 5G.

Also, if much more new competition is expected, or if significant portions of the addressable market were to shrink, the investment case might be more challenging than some propose.

It might be rational to “go slow” on 5G investments, if in fact one believes market consolidation is about to happen, and capital will have to be expended to buy assets. Conversely, if one expects to be a seller, then delaying an expensive investment also would make sense.

All of that makes even more sense if the near-term 5G revenue upside is expected to be muted.

Big Tier-One Telcos Face Huge Issues, Including Harvesting Versus Growth

It long has been a sound bit of advice for firms operating in declining industries to harvest revenue, rather than invest. AT&T did that in its long distance business, and now is doing so in its voice and messaging business, as well as its linear video business.

But what to do about fixed network internet access investments is a big issue in many markets where the fixed network business is big, but declining, or small and barely growing. That is especially true when more growth is expected in the mobile segment.

Many observers were critical of AT&T for buying DirecTV, and are critical of AT&T’s effort to buy Time Warner. Some of the criticisms are similar to complaints made when AT&T was considering the purchase of DirecTV.

Capital should instead have been investing in the core fixed line network, instead, many argued when AT&T was considering the acquisition of DirecTV.

Now the issue is whether AT&T should make access investments in place of buying Time Warner.

But it is not clear how much upside exists for AT&T, in terms of fixed network internet access revenue. You might argue that the best case for AT&T, for a massive upgrade of its consumer access network, is about 10 percent upside in terms of consumer market share.

That is by no means insignificant, depending on the assumptions one makes about the cost of the upgrades. Still, given that as important as it is, fixed network internet access now is a mature business, there are limits to how much capital a telco “ought” to invest, compared to deploying capital elsewhere.

Realistically, a major telco has to expect it will, under the best of circumstances, and in a two-provider market, split share with a competent and motivated cable TV provider. If cable now has about 60 percent share, and AT&T about 40 percent share, that implies a sort of share ceiling of 50 percent. That is one driver of revenue. The other is revenue per account.

But typical account revenues have not risen as much as one might expect, given consumer shifts to higher-speed services that tend to cost more.

Basically, internet access prices in the developed world have tended to move roughly in line with growth in gross domestic product, and are flat to declining in terms of spending as a percentage of gross national income per person, according to the International Telecommunications Union.  

source: PwC

Though some believe there is much-more account revenue upside, experience so far suggests prices will be hard to boost, in the fixed network segment. Increasing competition from third parties is one reason. A shift to mobile access is another issue. Finally, consumers will only spend so much on communications services in general.

At least some might point to stock performance of T-Mobile US, compare that to AT&T, and draw the conclusion that AT&T would be better off putting its capital into network upgrades, not content acquisitions.

That mistakes a growth strategy based on taking market share with a strategy based on entering or creating new markets. Even if markets are not growing, attacking firms can grow simply by taking market share.

That is what cable TV companies have done in business service markets, voice services and internet access, for example. Growing by taking share was not possible for AT&T and Verizon, which already were the market share leaders in those markets.

That is germane when looking at T-Mobile US strategy, compared to that of AT&T or Verizon.

For starters, T-Mobile US--with no fixed network footprint--has only one avenue for growth: taking market share from other mobile service providers, something it has done.

And, in a zero-sum U.S. mobile market, T-Mobile US, with market share of about 15 percent, has room to grow at the expense of the other service providers, until some future time when it will be acquired or merged with another sizable firm.

In fact, should current predictions about the 5G era prove correct, T-Mobile US and Sprint might well require major fixed network assets to support small cell networks.

The point is that T-Mobile US has limitations and opportunities in its core business that are quite different from those of AT&T. Neither Verizon nor AT&T has seen much share change over the last decade.

So what makes sense--and is doable--for T-Mobile US is not necessarily sensible--or doable--for AT&T or Verizon.

As voice and messaging already have entered the declining part of their product life cycles, so too internet access and mobility itself have reached near saturation in the U.S. market. That means finding brand new sources of revenue growth beyond the legacy core.

That is not to say AT&T or Verizon can afford to neglect upgrades of their consumer internet access capabilities. But it would also be incorrect to argue that such upgrades can drive overall revenues over a decade’s time.

One might argue that U.S. mobile internet access revenue will grow. It is harder to make the argument that fixed network revenue will do much, based on past experience.

source: PwC

AT&T is a firm with many big decisions to make, and none of them are especially easy. Where to deploy capital for revenue growth is a prime example.

Charter Plans to Use Consumer Modems as 4G/5G Radio Sites

Major cable TV companies long have been planning to use their networks as platforms to support mobile service, which is why Comcast home gateways have both the end user access and separate access for the “public” Wi-Fi function.

Charter is now talking about the next step, which is to add an actual 4G Long Term Evolution and 5G radio function, making the in-home router a small cell.

source: Charter Communications

What is the 5G Use Case?

Some things do not change. Consider the business model debates we once had about 4G and 3G, in relation to 5G. Most proponents said those new platforms (and 5G) would enable creation of new services. The revisionist view was that the new networks mostly would lower the cost per bit of supplying mobile data access.

Sometimes apparently-contradictory explanations can both be correct. Both 3G and 4G were able to supply lower cost per bit data access. But it also is the case that some new use cases did develop, for 3G and 4G, and presumably also will happen with 5G.

And there likely will be a progression, in that regard. If you remember 3G, you will recall that the “new apps” did not develop for some time. Mobile email access (BlackBerry) might seem a rather trivial matter today, but it drove a huge change in behavior and revenue in the 3G era, as did mobile web access.  

The same might be said about 4G, which eventually lead to massive use of streaming video, as well as tethering.

The immediate use case was in fact  “faster internet access,” even if mobile video was the new use case enabled by the faster access.  

For 5G, the same process is likely to unfold: faster internet access will be the immediate new capability customers use. The one immediate new use case will be fixed wireless, built on the 5G infrastructure, to be a full substitute for fixed access provided by cabled networks.

One might argue that represents a break from past practice (fixed wireless is a “new app or use case”). Or, one might argue fixed wireless is more an instance of “faster access” (compared to either digital subscriber line or mobile access) and hence more in keeping with past developments.

The larger point is that, initially, faster access (and lower cost per bit) will be the use case and business value.

Only later will entirely-new use cases (connected car, ultra-low-latency use cases, massive machine-to-machine apps) develop.

So it is a false dichotomy to position 5G either as a “lower cost per bit” value instead of a “platform for new apps” value. As has been the case with 3G and 4G, 5G will represent both sets of values.

As lower cost per bit has been a feature of each next generation mobile platform, and arguably also the key to enabling use of internet web, app and streaming use cases, 5G will do the same.

Is CBRS Wi-Fi Redux?

Every few decades, it seems, we have new debates and questions about whether a network access platform using unlicensed or low-cost spectrum can emerge as a full replacement for mobile network access.

A separate question might be whether networks using unlicensed spectrum or low-cost shared spectrum can become a full substitute for fixed networks built using cables, and if so, on what scale. On that score, fixed wireless networks have been operated for decades by wireless internet service providers using unlicensed spectrum.

The bigger question is whether local networks using unlicensed or low-cost shared spectrum can become replacements for wide area mobile networks.

Rethink Technology, for example, has argued that shared spectrum (Citizens Broadband Radio Service; use of unlicensed and shared spectrum) will trigger new “challengers” to mobile operator revenue models.

source: Rethink Technology Research

Caroline Gabriel, director of research at Rethink Technology Research, believes that “enterprise small cells, particularly those operating in unlicensed spectrum, could be the undoing of mobile network operators, relegating them to ‘utility’ commoditization, and falling revenues, as both license free spectrum and shared spectrum ideas, are taken up aggressively by newcomers.”

“Small cells were conceived as a way to improve the mobile operator business model, but they may now become weapons for challengers to MNOs, particularly broadband players with established backhaul such as cable operators,” she says.

Without question, use of unlicensed or shared spectrum to create connectivity services can affect the business model of some access service providers.

Service providers with a huge network footprint (cable TV operators being the best example), can use Wi-Fi access to reduce the amount of capacity they buy from their wholesale providers, when operating as mobile service providers. So, without question, unlicensed and shared spectrum will have business consequences for some competitors.

Independent wireless ISPs might also be able to leverage new shared and unlicensed spectrum.

In that sense, yes, new shared spectrum and unlicensed spectrum can--and will--be used by telco competitors, as well as by telcos themselves.

A related question is whether use of unlicensed or shared spectrum by large enterprises actually displaces access that otherwise would have been purchased from a mobile service provider.

That could be relevant in terms of internet of things deployments, where an enterprise might build its own local network to collect data from sensors, rather than using services from a mobile service provider, instead.

Rethink estimates that, by 2022, enterprises will account for almost half of all small cell deployments, up from seven percent in 2014. The installed base will reach 14.8 million in 2022, up from 185,000 in 2014.

The issue is why enterprises will be deploying, what the use cases are, and how such deployments could affect service provider revenue.

Here one must look to history.

Substitution of this type (private networks for public networks) has been talked about for decades. Wi-Fi provides the example. Back in the late 1990s and early 2000s, one could hear some talk of the theory that Wi-Fi eventually would emerge as a full access substitute for mobile network access.

That mostly happened in the early 3G era.

Conversely, perhaps some might argue that new mobile networks will obviate the need for Wi-Fi. That almost assuredly goes too far, but even Cisco’s forecasts waffle on this matter.

A 2015 forecast by Cisco suggested use of Wi-Fi would not decrease, as a percentage of total access. Later forecasts suggest 5G will increase the amount of access on the mobile network, reversing a trend in place since 2G.

source: Cisco

One potentially useful analogy is the traditional use cases for private, indoor networks (local area networks and Wi-Fi) and outdoor mobile and fixed networks. One can argue that the traditional bifurcation between public “access” networks and private enterprise networks still provides a likely model.

Wi-Fi essentially is the platform that replaced cabled private premises networks. Public carriers earned their revenue bringing connectivity to the premises, but enterprises and organizations and built their own internal distribution networks.

Rethink believes CBRS will therefore enable replacement of mobile services by enterprise owners building private networks. That could happen, in the sense of enterprises substituting private network access for public network sensor access.

But that would limit new public network IoT connections, not displace existing mobile connections. There would be a revenue impact, but of an indirect sort: eliminating one potential source of new mobile network connectivity revenues.

The analogy is not perfect, but the bifurcation between private local area networks and wide area public networks is germane. That is not to deny the emergence of “neutral host” providers selling wholesale premises access to all mobile service providers.

But, in such cases, the neutral host providers essentially are mobile infrastructure providers working in partnership with their mobile clients, as do tower companies.

We will wait to see how the business models develop, but LANs and tower company precedents, as well as services such as Boingo, provide some prior examples of how the new CBRS and related markets could develop.

Intel 5G Product Roadmap

It is a truism that next generation networks can be used by customers only when their handsets can support the protocols and frequencies used by those networks. So Intel has announced a 5G product roadmap that does so.

The Intel XMMTM 8000 series 5G new radio (5G NR) multi-mode commercial modems are part of the story, operating in sub-6 GHz and millimeter wave frequencies globally. The Intel XMMTM 8060, Intel’s first commercial 5G modem,  is capable of delivering multi-mode support for the full 5G non-standalone and standalone NR, as well as various 2G, 3G (including CDMA) and 4G legacy modes.

The latest LTE modem, the Intel XMMTM 7660, supports speeds up to 1.6 Gbps.

Intel also announced it has successfully completed a full end-to-end 5G call based on its early 5G silicon, the Intel 5G Modem, including use of 28 GHz frequencies.

Finally, the Intel XMMTM 7560 modem unveiled at Mobile World Congress 2017 has achieved gigabit-class speeds.