NFV, SDN, Cloud Native

Most of the time we talk about networks functions virtualization and software defined networks as ways to lower the cost of operating a modern network (the former) or enable new services (the latter). Some call this overall process a transition to "cloud native" communications. 

Open source also is among the many ways network operators are wringing cost out of their platforms. 

All those trends illustrate the ways that modern telecom networks have become computing networks. One big business model implication is that applications and services are logically separated on a modern computing network. That is why we hear so much about "dumb pipe" and "OTT." 

Will 5G Cause Economic Growth?

Correlation is not causation, it is helpful to remember. There is a correlation between higher per capita gross domestic product (GDP) and more time spent on Wi-Fi, say researchers at OpenSignal.

There also is a correlation between higher GDP and overall average mobile upload and download speeds; overall Wi-Fi download speeds; and LTE availability, OpenSignal researchers note.

Others will note that there is a correlation between GDP and fixed network internet access availability; between gross domestic product and internet access speed;  or between GDP and affordability; between 5G and economic activity.

source: IHS

To reiterate, correlation exists. Causation is another matter.

Does internet access cause growth, or does growth lead to better, faster, more affordable internet access? It is impossible to know.  

The same sorts of issues can be noted where it comes to Wi-Fi speed. There is a correlation between faster Wi-Fi (faster fixed network access) and GDP.

source: OpenSignal

But it remains impossible to determine the extent to which better connectivity causes growth, or the extent to which growth leads to better connectivity. Put another way, does wealth lead to spending, or does spending lead to wealth? Nobody doubts the correlation. But to the extent public policy is based on perceived causation, the causal chain does matter.

“Necessary but not sufficient” often is the best formulation. Quality internet access might be necessary for economic growth, but is not sufficient to cause it. Good communications might be necessary to attract firm foundings, relocations or expansions. Rarely, if ever,  is it sufficient.

Good internet access is an aid to quality education or prevent population loss. It is not sufficient to cause educational quality or prevent population loss.

Wi-Fi Offload Will be Replaced by Wi-Fi Bonding

Wi-Fi offload has been a central part of mobile connectivity strategy on the part of users and service providers since 3G. That has changed in the 4G era, and might flip in the 5G era, when it might make sense to stay on the mobile network all the time, tariffs permitting.

Already, in  33 countries smartphone users now experience faster average download speeds using a mobile network than using Wi-Fi, according to OpenSignal mobile analytics.

In 50 countries, 63 percent of those studied, 4G networks offer a faster smartphone download experience than Wi-Fi.

Equivalent download experience means there is no “speed reason” for smartphone users to switch network type. But there still might be tariff reasons for doing so (mobile data is expensive, compared to Wi-Fi, either public or private).

In a number of markets with well-developed fixed network access, Wi-Fi offload might still be faster than 4G, though. But it is no longer automatically the case that Wi-Fi offload provides a better experience.

source: OpenSignal

So mobile operators and smartphone makers must re-evaluate their Wi-Fi strategies, especially around mobile offload, automatic network selection and indoor coverage, to ensure they do not accidentally push consumers’ smartphones onto a Wi-Fi network with a worse experience than the mobile network, researchers at OpenSignal say.

source: OpenSignal

Can IoT Edge Computing be Leveraged for Connecting the Unconnected?

Many of the use cases cited for edge computing also match with conditions of human users in rural and isolated areas, suggesting that we might find ways to leverage advanced market edge computing tools to support computing in remote and rural areas as well.

Consider a few of the characteristics of situations where edge computing is useful, because of:

• Intermittent connectivity (communications network unreliable)
• Localized compute power (unreliable connectivity)
• Intermittent power supply (power grid unreliable)
• Rugged environment (jungle, desert, arctic, mountains)

The one situational characteristic that is not absolutely relevant is real-time decision-making.

source: GSMA

The point is that computing architectures are shifting again, from centralized to distributed, in a way that has not generally been the case since the advent of cloud computing. Depending on how computing at the edge is done, we are shifting back to computing by the edge device (PCs with locally-resident apps on the hard drive) in some cases.

In other cases we are shifting back to a client-server model, with user devices relying on a local server in the office or on the factory floor.

The issue is how much network features and devices developed for internet of things apps under conditions of intermittent power and connectivity and more-rugged environments might be applied to computing in rural and hard-to-reach areas.

The biggest issue is that so much of human computing now relies on constant interaction with cloud-based servers. Caching has been a solution for such intermittent connection issues. The new possibility is that techniques developed to support IoT might be helpful for human computing in remote areas as well.

5G Fixed Wireless Arbitrage in U.K. Market

Arbitrage long has been a business model issue for access services, as illustrated by one advantage facilities-based 5G fixed wireless has over use of the Openreach platform.

According to researchers at Ovum: there is no need for consumers to pay £240 (about US$308) a year on line rental charges required when they buy any fixed network service offered using Openreach facilities.

5G fixed wireless, using millimeter wave spectrum, has performance is comparable to, or better, than existing fiber-based products, Ovum researchers say.

"We expect large-scale deployment to be able to consistently support speeds of 80 Mbps to 100 Mbps,” which is better than many existing fiber-to-curb connections, say researchers at Ovum. The average U.K. fixed network speed today is 46.2 Mbps.

Also, 5G-FWA can serve 85 percent of the existing UK fixed-line market, Ovum believes, based on current speed experiences.  The percentage of urban customers receiving speeds below 80 Mbps is approximately 85 percent.

Overall, Ovum estimates 5G fixed wireless is almost 50 percent cheaper to deploy than fiber to premises, for example, based largely on construction costs. Nokia sees a positive business case when a fixed wireless network can reach 30 customers served per cell site, though obviously the model is sensitive to capital investment and average revenue per account.

The business model requires a 30-percent take rate. That is a lower threshold than generally required for a facilities-based fiber-to-home network.

Average revenue per account probably needs to be at least 40€ (US$43) per month, a target most consumer service providers should be able to hit with one service.

source: Nokia

Subscription Video is Too Big to Ignore

Telco TV (linear video delivered over the fixed network) has taken a bigger beating over the last few years, as the whole sector experiences contraction. Still, entertainment video is among the few mass market consumer services ever to have developed. Voice services and internet access are the only other two services with near-ubiquitous demand, on a fixed network.

And with fixed network voice lines in serious decline, and market share in the internet access market severely training cable TV operator share, entertainment video is the only other revenue driver that can affect overall revenue, even as it declines.

In 2018, linear video represents about $126 billion in annual revenues. Fixed network internet access was about at least a $60 billion business in 2018. If cable operators have 65 percent of that, then telco internet access generates about $21 billion annually.  

Keep in mind that total U.S fixed network telco revenues were only about $75 billion in total, in 2018. Assume 40 percent is generated by business customer services, leaving about $45 billion total that is generated by consumers. That suggest voice generates $24 billion a year in revenue.

The point is that video entertainment is a much-bigger business than voice or internet access, in the fixed network segment of the U.S. business.

source: UBS

source: S&P Global Market Intelligence

source: Strategy Analytics

How Big is the U.S. Business Market?

How big is the U.S. business customer market for connectivity services?

AT&T earned about $7 billion in fixed network business revenue in its third quarter of 2018. Make annual revenue $28 billion. Total business segment sales, including mobility services, was about $9.2 billion for the quarter. So make annual business segment revenue $36.8 billion.

Verizon earned about $4.2 billion in fixed network business revenues in the third quarter. Make that $16.9 billion on an annualized basis. Verizon also earned about $16 billion in the third quarter in mobile revenues. Assume 14 percent of that is business customer revenues, or $2.2 billion. Annualize business mobile revenues to $9 billion. That in turn leads to annual business customer revenues of about $26 billion.

CenturyLink booked about $3.4 billion in U.S. business revenues. Make that $13.6 billion in annual business revenues. Based solely on sales by those three companies, U.S. businesses spend $76 billion on connectivity services. Add $18 billion in revenue earned by cable operators to that total and one derives $94.4 billion in annual business customer connectivity revenues.

Then add about 20 percent more revenue earned by all other service providers in the “telco” segment (20 percent of $76 billion), to add $15.2 billion. The grand total for business connectivity services then would be about $109 billion.

That is quite a bit lower than many other estimates.

By some estimates, U.S. business and government spending on communications is about $256 billion annually.

By other estimates, total U.S. telecom spending is about $376 billion annually. As a rule of thumb, business spending drives 40 percent at minimum to perhaps half of all telecom spending. So that would imply business connectivity spending (not including phones and other equipment) is between $150 billion and $188 billion.

By such measures, U.S. cable operators have gotten between 9.5 percent and 12 percent share of the U.S. business connectivity revenues market.

source: Accedian

Another way to estimate U.S. business spending is to assume that U.S. telecom spending is about 25 percent of global telecom spending, which is perhaps $1.5 trillion. That implies U.S. spending of perhaps $300 billion. If business spending is 40 percent of that, one derives a figure of $128 billion. If business spending is as much as half of total U.S. connectivity spending, the figure could be $150 billion.

The point is that a conservative metric, the U.S. cable TV industry already has gotten as much as 16.5 percent share of the business communications market.