Hard Work, Moore's Law, Human Cleverness for Breathing New Life into Many Legacy Platforms

At one point a couple of decades ago, even sophisticated access technologists might privately have thought standard twisted-pair cables used by telcos would not work reliably enough to deliver 10 Mbps to scores of megabits per second.

By the same token, many would not then have believed cable TV hybrid fiber coax networks capable of reaching hundreds of megabits per second, much less gigabit speeds. But credit hard work, technical cleverness and Moore’s Law for breaking all those limits.

Swisscom is deploying G.fast, which allows Swisscom to reach transmission speeds of up to 500 Mbps on networks with relatively short copper drop cables and optical fiber distribution networks.

That deployment is part of a process whereby legacy networks are upgraded to deliver bandwidth that once was thought “impossible” on telco legacy copper drops, cable TV hybrid fiber coax, fixed wireless or mobile networks

Better radios, signal processing and advanced modulation techniques, plus new architectures, are also are contributing to the way mobile and fixed wireless networks is improving

In the medium to long term, Swisscom intends to supply 85 percent of all Swiss households and businesses with bandwidth of at least 100 Mbps by the end of 2020.

Sprint Makes Progress

Long-ailing Sprint seems to be making operational progress, growing operating income, adding postpaid accounts and reducing churn.

Unaudited Sprint preliminary financial results for its second fiscal quarter of 2016 (three months ended September 30, 2016) show progress.

Total net operating revenues of $8.25 billion grew three percent year-over-year and wireless net operating revenues of $7.85 billion grew nearly five percent year-over-year.

Sprint said it lost $142 million in the third quarter, compared with a loss of $585 million in the same months of 2015. Operations profited by $622 million, though the operating profit total ignores some costs such as debt repayments or interest payments

The company added 740,000 subscribers during July, August and September, including 344,000 in its higher-revenue customer counts. Sprint reduced postpaid churn, which dipped from 1.54 percent last year to 1.52 percent this year. The carrier noted postpaid phone churn for the latest quarter hit a company record low of 1.37 percent.

With T-Mobile US driving its financial and operating results and taking market share in the U.S. mobile market, and Sprint doing better, one has to wonder if AT&T and Verizon will show a bit of weakness in their next quarterly reporting.

Beyond that, there are the coming market entries of Comcast and Charter Communications, as well as the unknown strategy Dish Network will employ, either entering the mobile market or selling its spectrum assets.

One way or the other, further turbulence lies ahead.

Top US Wireless Carrier Metrics Q2 2016 (ranking by subscribers, retail + wholesale)
Rank by             Subscribers	Carrier	Subscribers (millions)	Net Adds (millions)	Postpaid Smartphone Net Adds (mil)
1	Verizo n Wireless	142.754	1.285	0.336
2	AT&T	131.805	1.361	0.250
3	T-Mobile USA	67.384	1.881	0.877
4	Sprint	58.446	-0.360	0.259

source: Jackdaw Research

Is Spectrum Value Growing, Flat or Declining?

Is the value of spectrum in the U.S. market flat, increasing or declining? It’s a hard question to answer in the abstract.

It all depends on which frequencies we are looking at, the specific value to potential owners who might buy spectrum, and other issues such as the amount of spectrum, the contiguity of the spectrum and the other alternatives which might be available to provide desired spectrum assets.

It might be fair to say there are big disconnects, in some cases.

Some have estimated the total value of 2.5-GHz spectrum held by Sprint at $115 billion or so. Others might argue that all of Sprint’s spectrum is worth around $60 billion.

It is worth noting both those figures exceed Sprint’s total market valuation of about $28 billion, in the third week of October 2016. Somebody is wrong, by quite a lot.

“We estimate that Sprint is valuing principally 2.5 GHz spectrum at $1.85/MHz/POP, which is more than six times the $0.30/MHz/POP it effectively paid for this spectrum when it acquired Clearwire in 2013,” says BTIG equity analyst Walter Piecyk.

How much is Dish Network spectrum worth? Observers continue to argue about the matter.  

Analysts at Kerrisdale Capital have argued that demand for Dish Network’s spectrum is wildly optimistic, and that spectrum prices are headed dramatically lower.

Others have argued that the spectrum represents most of the equity value of Dish Network as a whole. Some have pegged the mobile spectrum licenses as 80 percent of Dish equity value, based on a valuation of $35 billion to $50 billion for the spectrum licenses.

By some estimates, facilities-based U.S. mobile operators, plus Dish Network, own about $368 billion worth of spectrum licenses.

AT&T now holds spectrum licenses worth more than $91 billion, estimates Goldman Sachs analyst Brett Feldman, while the value of Verizon’s spectrum is $79.4 billion.

In all, AT&T now holds spectrum licenses worth more than $91 billion, estimates Goldman Sachs analyst Brett Feldman. He also estimates the value of Verizon's spectrum at $79.4 billion.

The current equity value of all AT&T stock is $176.5 billion, implying that spectrum alone represents 51.6 percent of AT&T’s total equity value.

Verizon’s market value is $207.9 billion, implying that Verizon’s spectrum represents 38 percent of total valuation.  

Bloomberg Intelligence has estimated the total value of Sprint’s 2.5-GHz spectrum alone at $115.1, about 2.4 times Sprint’s enterprise value of $48 billion.

In fact, some have argued that T-Mobile US spectrum accounts for more than 100 percent of its total market value.

Sprint apparently values 14 percent of its spectrum holdings at $16.4 billion as part of a recent sale-leaseback of spectrum.

That implies a total valuation of spectrum at about $117 billion, or about four times Sprint’s present market capitalization. Clearly there is a huge variance; some might say a disconnect.

Either Sprint’s spectrum is not worth as much as it claims, or the market is seriously undervaluing Sprint as an asset.

Dish Network has a huge interest in spectrum valuation , as it holds a significant block of mobile spectrum that must either be put to use, or sold, or returned to the government.

“If we valued Dish’s core business at five times the consensus 2016 EBITDA estimate and $1.85 per MHz per POP it would imply a value of $106 per share for Dish and $85 per share if we taxed the gains from a sale of spectrum at that level at 35 percent,” Piecyk argues.

In the third week of October 2016, Dish Network’s equity is selling for about $57 a share. Again, there is a disconnect between implicit spectrum value and equity value of the whole business using--or potentially using--that spectrum.

So the value of mobile spectrum matters especially for a few firms (Dish Network and Sprint, in particular) that are monetizing, or hoping to monetize, those assets.

The specific value of existing spectrum matters less to AT&T, Verizon and T-Mobile US, as it is an asset meant only to support the business model, and not being collateralized (Sprint) or potentially monetized (Dish Network).

Will Amazon Become an ISP?

Ironically, as Amazon reportedly ponders becoming a retail Internet service provider in the United Kingdom or Germany, the broad implications of mandatory wholesale again come into perspective.

The attraction for Amazon of becoming an ISP: offering a bundle of Internet access and Prime video, essentially making Amazon a dual-play competitor to both telcos and cable companies.

The point arguably is to boost Prime uptake, more than the attractions of revenue earned by supplying Internet access.

Such a move, already taken by Google Fiber in the United States, has not been directly matched by Facebook, which is focusing more on backhaul platforms that work with mobile operator efforts. Google’s Project Loon presently contemplates a backhaul role, with mobile operators being the retail access providers.

The broader background issue is the role of wholesale access in the access business, which varies by country. On one hand, the robust wholesale regime allows many retailers to enter a market.

About 80 percent of retail U.K Internet access accounts are supplied using BT’s wholesale network. About 20 percent is supplied by Virgin Mobile, using its own network.

In Germany, perhaps 17 percent of total Internet access connections in early 2015 were supplied by cable networks.

In France, where most fixed network Internet access connections happen over the Orange network, using wholesale mechanisms, perhaps four percent are supplied on cable TV networks.

The point is that most retail ISPs, building their business on wholesale access, make a trade off. They gain the ability to supply access, without building their own expensive facilities, at the price of losing control of their access costs and features.

source: ARCEP

As a practical matter, that means every retail ISP using BT’s wholesale network is limited to the speeds and features available to every retail service provider using the network.

Talk Talk, which is building (in some instances) its own fiber to home access connections, offers service at 900 Mbps. Though not yet a widespread option, operating its own facilities allows Talk Talk to vastly outperform the typical U.K. access connection.

Though Virgin Media is upgrading, “average” speeds for most fiber-to-home connections runs about 42 Mbps, or 54 Mbps for cable connections.

source: Ofcom

A major disadvantage obviously is that no retail provider is able to distinguish itself on network features, and each supplier’s cost structure is bounded the cost of leasing the wholesale assets.

Virgin Mobile is able to offer a differentiated service because it runs over its own network, indeed using a different platform (hybrid fiber coax, not fiber to neighborhood or fiber to home).

There always are trade offs in the communications business. For many retail ISPs operating in the fixed networks business, the decision is between lower capex, without control; or high capex, but with ability to differentiate.

Amazon might well deem the ability to operate at lower capex the key value, as Amazon Prime video does not require excessive bandwidth.

In Asia, Users Sometimes Use Apps More than Web; In Other Cases Use Web More than Apps

Global statistics can obscure more than they reveal, and that is true for the Asia-Pacific region as well.

Asia will account for 55 percent of the world’s nearly 1.18 billion smartphone users in 2016, according to eMarketer. But user behavior is quite disparate, where it comes to preferences for use of mobile web and apps.

In more developed markets like Japan and South Korea, daily time spent with the mobile web on smartphones outpaced time spent with smartphone apps, often by 10 minutes or more, a TNS study found.

In emerging markets in Southeast Asia including the Philippines, Vietnam and Thailand, daily time spent was more likely to favor smartphone apps.

This split in app habits follows mobile trends observed in other regions of the world. Mobile users in emerging markets are comparatively more enthusiastic about downloading and spending time with apps, those in more developed economies less so.

Telcos Upgrade Networks In Wealthier Areas, Lag in Poorer Areas, Study Argues

In a study, the Center for Public Integrity argues “the largest non-cable internet providers collectively offer faster speeds to about 40 percent of the population they serve nationwide in wealthy areas compared with just 22 percent of the population in poor areas.”

The argument is that the large telcos have rather systematically targeted network upgrades to “wealthier” communities, compared to “poorer” communities.

There is statistical truth to those claims. Any analysis of urban-rural Internet access speeds would show that rural speeds are generally much slower than urban speeds, and that urban household incomes are generally higher than rural incomes.

Hence, it is is statistically correct to argue that speeds are, in fact, faster in urban (wealthier) areas, compared to rural (poorer) areas.

It might also be correct to argue that top speeds in some poorer neighborhoods are slower than top speeds in more-upscale neighborhoods.

The reality is that several obvious issues are at work. Telco copper network speeds are distance sensitive. That is one reason why rural networks have been “slower” than urban networks, as a rule.

Customer density is far lower in rural areas, which is why it is hard to justify the same investment in rural areas, compared to urban areas.

source: FCC

Also, upgrading even some urban properties--especially multiple dwelling units--is less than straightforward, as building owners can block or delay installation of new facilities.

Beyond all that, communications policy related to Internet access facilities never has been governed by “common carrier” regulation that mandates universal access and comparable prices for the same classes of service, or even minimum levels of service.

In fact, precisely because incentives matter, where it comes to next-generation facilities, mobile and fixed service providers have traditionally had more leeway to build facilities in some areas first, in other areas later, to avoid building, in many cases.

Service providers are free, for example, to build new networks to serve business customers only, and not consumers; large businesses in preference to small businesses; or businesses in some areas and not others.

In fact, municipal authorities now have taken that same approach and applied it to next generation networks serving consumers. That is why Google Fiber is allowed to build in some neighborhoods, and not others, or to build first in some neighborhoods, and not others.

That approach has changed AT&T and Verizon thinking about deployment of fiber to home and gigabit networks, as well. Building first where there is higher demand drives more investment. It also means disparities will widen, for a time.

So, yes, the CPI study does show disparities. But the explanation is partly for reasons of regulatory frameworks, incentives for investment, the physical properties of telco copper networks, population density, property owner rights and end user demand for services, as well as income disparities.

That is not to condone the disparities. But it is a reality of the business model that networks are far more expensive in rural areas than urban areas, while demand is more robust in business customer segments than consumer; higher in wealthier neighborhoods than poorer ones.

There likely are disparities by age, marital status, children in household and regional differences, as well.

As this analysis by the FCC shows, high population density means affordable network cost. Conversely, low population density means high network cost.

It No Longer Really Matters Whether "Users Need a Gig"

It probably no longer makes sense to argue about whether gigabit Internet access “should” be deployed at a time when “nobody needs a gig.” The reason is that the shift to gigabit networks is not driven by direct end user demand, at the moment.

Instead, gigabit access speeds are strategic matter, driven by the level of intensifying competition in fixed and mobile networks, all of which are moving rather swiftly in the direction of gigabit and multi-gigabit capabilities. And that is being done irrespective of immediate end user demand.

Australia’s National Broadband Network (NBN), for example, says “14 percent of NBN’s fixed line services used a 100/40 Mbps wholesale speed tier, 49 percent used a 25/5 Mbps wholesale speed tier and 32 percent used a 12/1 Mbps wholesale speed tier” in 2015.

In 2014, those tiers were at 18 percent for 100 Mbps plans, 42 percent for 25 Mbps plans and 35 percent for 12 Mbps plans.

Note that demand for 100 Mbps actually dropped, year over year, while demand for 25 Mbps plans grew.

It is one thing to argue that the cost of upgrading immediately to gigabit fiber-to-home networks will not yield sufficient return to justify making the investment.

It is quite another matter to argue that cable TV operators, using hybrid fiber coax, likely can justify selling gigabit speeds, and already are preparing for multi-gigabit speeds, over their hybrid networks, without resorting to fiber to the home.

Likewise, the global mobile ecosystem is preparing for “gigabit to every device” in the earliest versions of 5G networks, with clear development paths for multi-gigabit and 10 Gbps 5G speeds.

At the same time, 4G networks are using channel bonding to boost 4G speeds up to the multi-hundreds of megabits per second range, whether or not there is direct and immediate end user demand.

As has been the case in the past, next generation networks are being developed for strategic reasons, not driven by direct end user demand.

Some competitors not required to use fiber to home platforms are going to market gigabit services themselves, in large part to maintain or take a lead over their key competitors. Mobile operators, on the other hand, see an opportunity to take market share from fixed services, for the first time, on a widespread scale.

Under such conditions, it doesn’t make much sense to debate, any longer, whether “users need gigabit.” They do not, at present. That is not the point, however. Internet service providers see a need to provide such capabilities, for strategic reasons and as clear marketing tools.

AT&T Uses Drones to Check Cell Towers

AT&T is using drones to conduct cell tower inspections and find birds' nests that can potentially affect cell coverage. That involves perhaps thousands of technicians every year. But that is going to change, if AT&T gets its way.

Eventually, AT&T says it plans to use artificial-intelligence-equipped drones to help assess problems in the field autonomously, without requiring a technician to manually inspect a tower.

As with all important new technologies, there will be job impact. Without a drone, inspecting a bird's nest could take up to a week and often requires the assistance of an environmental scientist.

The specific use of drones to inspect towers without sending techs climbing up those towers will save on operating costs, part of AT&T’s wider effort to slice such costs as part of its aggressive software defined network initiative.

It will not be popular, but lots of other measures will have to be taken to streamline network capex and opex, to match expected costs with expected revenues. And with competition seemingly growing, revenue is going to get hit. That means costs have to fall.

And one person’s cost is another person’s income. There’s no way around that.

Study Suggests 3 Facilities-Based Competitors Unlikely to Succeed in U.K. Market

source: Analysys Mason

A new analysis of facilities-based competition in the United Kingdom tends to support the view that ubiquitous fixed network on a facilities basis will tend to be a duopoly; that a third network offering triple-play services would not be able to cover most of the country, even under the most-optimistic scenarios.

That conclusion likely comports with the notions most researchers have reached, namely that facilities-based and ubiquitous fixed access network markets are oligopolies, because that is all the market will tend to support.

The only issue here is whether the stable oligopoly structure (in terms of facilities) features two providers or can include three (and if so, under what conditions).

That is not to say an oligopoly at the retail level is inevitable: robust wholesale policies have proven effective at stimulating retail competition. What robust wholesale has tended not to promote is investment in new facilities.

Analysys Mason researchers estimate that, at about 25 percent market share, a third facilities-based provider would be able to sustain access to about seven percent (two million locations) of homes.

The study tends to reconfirm that housing density encourages investment. But the study also suggests that robust wholesale policies actually discourage investment.

source: Analysys Mason

The researchers note that the greatest degree of facilities-based competition elsewhere in Europe relies on duct and pole access by competitors in countries where there is no wholesale regime in place.

In other words, where competitors must build their own facilities, and have some clear incentives to do so, facilities-based investment is most robust.

Also, third parties seem to have been most successful where the percentage of high-density housing is high.

The study authors conclude that investment in a third facilities-based network is highly risky and unlikely to succeed in the U.K. market, on a wide basis, though the business case might work in some high-density areas representing about 1.1 million households.

The analysis concluded that "it is highly unlikely that a third operator will be able to reach 40 percent (FTTP) coverage on a commercially viable basis".

“Our economic modelling suggests that encouraging a third separate network to invest in covering more than five to 10 percent of the country will be extremely difficult to achieve,” the study states.

Some might point to the key role cable TV competition tends to play, as well. In several countries, the primary facilities-based competition comes from cable TV operators.

For All Networks, "Fiber to Where You Can Make Money" is the Issue

"Fiber to where you can make money" is a good way to evaluate various fixed network access methods. For cable TV operators, the issue is fiber deep into the neighborhood. For some telcos, that is the same issue.

For some telcos, fiber to the premises is the choice, but the issue of revenue generated by such networks remains.

For fixed wireless networks, the issue is fiber to tower or building. For mobile operators, the issue is fiber to the macrocell or fiber to the small cell.

In other words, the issue is not the choice of physical media, or the topology, but the revenue that any given deployment can generate. And, at a time when voice revenues are declining, and where facilities-based competition exists, the financial returns from fiber to the home often are questionable.

The choices are even more difficult for any telco operating in a market where cable TV firms are active, serious competitors. Stranded assets then are the real issue, as up to 60 percent of deployed assets can routinely be stranded (assuming cable TV gets 40 percent to 45 percent share, the telco gets 40 percent share, while other suppliers get 15 percent to 20 percent market share.

In other words, the issue is to deploy “fiber to where you make money.” That is true for all fixed network suppliers, including cable TV firms using hybrid fiber coax technology.

Assuming fiber is deployed rather deep into the network, coaxial cable can carry huge amounts of bandwidth.