Saturday, September 30, 2017

IoT, AI, Edge Computing, 5G, Fiber-Deep Networks All Related

The internet of things, artificial (or augmented) intelligence, edge computing, low-latency application requirements, 5G, fiber-deep networks and mobile small cell architectures all are related trends. IoT, in many cases, will require low-latency computing, using advanced AI techniques with processing necessarily localized.

That, in turn, dovetails with the use of 5G and other communication networks that will have to support low-latency communications and highly-distributed computing nodes using 5G air interfaces, small cell architectures and fiber-deep distribution networks.

source: CB Insights

Friday, September 29, 2017

Google Launches IoT Platform

One example of the challenges telcos will face as they ponder moving into platform portions of the internet of things ecosystem are moves by rivals with business advantages. Consider that Google has officially launched its Google Cloud IoT Core service into public beta.

Google Cloud IoT Core is a fully managed service on Google Cloud Platform (GCP) that helps businesses secure and manage their connected IoT solutions.

Google Cloud IoT Core integrates with other Google analytics services such as Google Cloud Pub/Sub, Google Cloud Dataflow, Google Cloud Bigtable, Google BigQuery, and Google Cloud Machine Learning Engine.

Pricing is based on the volume of data exchanged with Cloud IoT Core.


Rate of Cord Cutting Slowing?

“The number of current pay-TV customers who plan to cut the cord has actually declined, and the number of hours spent watching old-fashioned, time-slot television is growing,” said Peter Cunningham, Technology, Media, and Telecommunications Practice Lead at J.D. Power.

To be clear, a new J.D. Power survey only shows that the rate of decline has slowed. The percentage of customers who say they plan to cut the cord on pay-TV during the next 12 months has declined to eight percent this year from nine percent in 2016, the company says.


The study might be interpreted as suggesting streaming and linear viewing modes are reaching some sort of equilibrium. We will have to wait and see. Many other forecasts suggest that newer modes (mobile, especially) are growing, though that does not directly speak to the issue of linear versus on-demand viewing.

Despite growing satisfaction with streaming video services and widespread use of DVR and video on-demand, the number of hours spent watching regularly scheduled television programs has increased by nearly an hour between 2015 and 2017, J.D. Power says.

In a typical week, households have spent an average of 17.4 hours watching regularly scheduled programming in 2017, up from 16.6 in 2015.



Thursday, September 28, 2017

Why Voice is Not Central for Next Generation Networks

A fateful decision was made when the global telecom industry decided the next generation network would be based on Internet Protocol, and not some traditional architecture such as ATM.

All of today's strategic issues around "over the top" services and value grow directly from that architectural choice, since IP fundamentally separates network access and all applications that run over the IP networks.

By definition, all apps are "over the top," no matter who owns those assets. In choosing to build the next generation networks on IP, the industry also chose to create the "dumb pipe" business model.

Even when telcos sell carrier voice, those services architecturally will operate "over the top."

At the same time, voice is receding as the core revenue driver.

According to Reza Arefi, Intel director of spectrum strategy, nobody is working on voice as part of 5G at any of the core standards bodies. That might come as a shock to many observers, but simply seems to point to the changing value of various revenue streams in the access business, and the fundamental way applications are created and delivered on modern networks.


The lack of focus on voice also is a reflection of changes in the core requirements for modern communications networks, where the growing range of capabilities come in the “connecting computing devices” area, not voice or messaging.


Also, the growing reality is that voice is a feature, less a key revenue driver. It is a key function, to be sure; just not the driver of revenue growth.


That, in part, explains the lack of work on voice as a core feature of 5G. That “neglect” is not new. You might recall that the 4G standard also did not originally support voice, either.


It might be reasonable to argue that 5G standards work does not include voice support because voice is seen as a service supported on 4G. Others might argue some extension of voice over Wi-Fi will be part of the solution.


Consider revenue drivers for the industry globally, which are predicted by STL Partners to continue declining significantly.
source: STL Partners

Nearly 1/2 of Linear Video Customers Might Consider Switching Next 6 Months

source: Tivo
Nearly 84 percent of respondents to the most-recent Tivo survey of consumers still buy a linear TV service, while 16 percent of respondents do not. “Price” was cited by 85 percent of the non-buyers as the reason for not subscribing. About 46 percent reported using a streaming service as an alternative.

Even if the top reason for not buying a video subscription is “price,” the actual reason likely is “value,” including both demand for video and its price.

Of the 83.9 percent of respondents who subscribe to pay-TV service, 8.9 percent have switched providers in the last three months, Tivo says. When asked if they planned to leave their current service providers in the next six months, 6.3 percent said they plan to cut their pay-TV service entirely.

Another  8.1 percent plan to change to another provider. In addition, 4.5 percent plan to switch to an online service or app. About 30.6 percent report they might make a change.

Altogether, 49.5 percent of respondents could potentially leave their current entertainment video provider in the next six months, Tivo says.

Fully 56.1 percent of respondents report they would stay with their current provider if they could choose and pay for only the channels they typically watch. Another 38.8 percent would stay if all video providers (Netflix, Hulu, Amazon Video and so forth) were combined into one place.

IoT Awareness Still Low

Many executives, in many industries, might be unaware of internet of things business value, a survey by Analysys Mason suggests.

In a survey of 1600 IT and telecoms decision makers in enterprises worldwide, Analysys Mason found 12 percent of small/medium businesses and 18 percent of large enterprises already had IoT systems operating.

On the other hand, 52 percent of SMEs and 40 percent of large enterprises were “unaware of IoT or not interested in it,” Analysys Mason found.

“Overcoming a lack of awareness in IoT is more important than technology issues,” Analysys Mason suggests.

None of those findings would surprise many observers. Present use of sensor networks is common in some industry verticals and for some applications, but it would be reasonable to argue that the big growth has yet to begin.


                     Percentage of enterprises at each stage of IoT development, 2017
igure 1: Percentage of enterprises at each stage of IoT development, 2017

The U.S. market currently has the highest levels of adoption of IoT solutions, but China is expected to surge.

              Percentage of enterprises with operational IoT solutions, by country/region, 2017
Figure 2: Percentage of enterprises with operational IoT solutions, by country/region, 2017



Wednesday, September 27, 2017

Enterprises Might Drive Incremental 5G Revenue

source: Bell Labs
If you work in the telecom industry long enough, you are likely to discover yourself changing focus from time to time. Having started out in cable TV, consumers were really 100 percent of my focus. When I switched to the competitive side of telecom, enterprises, wholesale and small and mid-size businesses became the focus.

When the internet hit, it was clear that consumer apps were going to drive growth and change, so I switched back to following consumer apps, business models and revenue streams almost exclusively.

I’m pretty sure that another shift back to enterprise is coming, since 5G is likely to be about enterprise, in terms of business success, incremental revenue and growth. Internet of things is the driver, since IoT will be “purchased” by enterprises and businesses, either to serve business users or consumers.

Some other inklings are that even enterprise wireless traffic looks to be on the cusp of significant change, according to Bell Labs.

Looking only at video, which is driving bandwidth requirements globally, there might be a shift there as well. Where today video is driven by consumer content and video, in the future it is possible that video will be driven by IoT devices.

Those are huge shifts. This time around, though, the emphasis likely will be on vertical industry shifts to IoT, not the particularities of access platforms.


source: Bell Labs

What Takes the Place of Mobility Over the Next Decade?

One fundamental rule I use when analyzing telecom service provider business models is to assume that half of current revenue has to be replaced every decade. One example is the change in composition of Verizon revenue between 1999 and 2013. In 1999, 82 percent of revenue was earned from the fixed network.

By 2013, 68 percent of revenue was earned by the mobile network. The same sort of change happened with cash flow (“earnings”). In 1999, the fixed network produced 82 percent of cash flow. By 2013, mobility was producing 89 percent of cash flow. The fixed network was creating only 11 percent of cash flow.


The picture at AT&T was similar. In 2000, AT&T earned 81 percent of revenue from fixed network services. By 2013, AT&T was earning 54 percent of total revenue from mobility services.


Also, consider CenturyLink. In 2017 (assuming the acquisition of Level 3 Communications is approved), CenturyLink will earn at least 76 percent of revenue from business customers. In the past, CenturyLink, like other rural carriers, earned most of its money from consumer accounts.

The point is that CenturyLink now is unusually positioned with respect to business revenue, earning a far greater percentage of total revenue from enterprise, small or mid-sized businesses and wholesale services, compared to other major providers.

After the combination with Level 3, CenturyLink will earn no more than 24 percent of total revenue from all consumer sources, and that contribution is likely to keep shrinking.

The big strategic issue is how revenue drivers will change over the next decade. As impossible as it seems, today’s mobility services are not likely to produce half of total revenues in a decade.

Up the Stack or Forward in the Value Chain?

Sometimes, moving up the stack is a business strategy, as hard as it is to achieve. More often, horizontal acquisitions for scale have driven telecom service provider acquisitions.

Sometimes, moving forward into the ecosystem supply chain also can work. Paradoxically, moving forward in the video entertainment subscription business (often understand as integrating lower inputs into a finished product or service) also means moving up the stack.

Either approach--up the stack or forward integration--is risky, in part because such moves represent moves outside the present understood core competence, but also because other firms will tend to resist buying horizontal functions and services from a key competitor.

Moving up the stack sometimes can mean acquiring assets that are a foundation for today’s core offerings. That paradoxical move is not, strictly speaking, a move “down the stack,” but is an example of vertically integrating a “cost of goods” input.

The best examples likely are video content producers. Owners of networks provide the core value behind the purchase of streaming video services. Strictly speaking, they operate up the stack.

Functionally, they are the basis for service provider video entertainment offers, a cost of goods that is almost a “down the stack” input, in a business sense. In other words, a video service cannot be created without access to the content consumers want to buy. In that sense, content is a raw material needed to build a video entertainment service.



Most telecom acquisitions have been of the horizontal variety, where a company buys additional assets in its existing place within the ecosystem (more access assets). That is an example of growing scale within the same business, instead of growing scope by occupying new roles within any ecosystem.


You might argue that acquiring such assets, that are required horizontally for all suppliers in a market, are a logical way for access providers to move up the stack. It never is easy, for simple reasons.

When making a horizontal acquisition, the acquiring firm simply gets bigger, doing what it already does. That means the acquiring firm already understands the business, and can hope to take out costs.

Moving into a different role within the ecosystem means moving outside the area of believed core competence. That always carries more business risk.
v
Also, when a function, or set of assets, is required by all service providers in a particular market, those competitors will logically try and avoid using a set of horizontal assets owned by a key competitor. So it is harder to occupy a new horizontal role. New business competencies are required, but competitors also will try and escape using the horizontal platform, if they can.

Big Questions about IoT Connection Choices

No matter how one looks at the 5G business model, there are huge questions. Consider the fundamental issue of how big incremental new revenue opportunities will be. The "big three" use cases always are said to be enhanced mobile broadband; low latency apps and massive machine-to-machine apps.

Briefly, here are the unknowns. Much of the 5G upside from enhanced mobile broadband initially will come from users upgrading from 4G. There is some incremental revenue upside, but most of the revenue will simply come from users switching from 4G to 5G.

In the IoT area (massive machine-to-machine apps), there are questions about which networks will gain. Not only are multiple network platforms conceivable, ranging form Wi-Fi and other local access options, but also dedicated IoT networks (LP-WAN), as well as 4G and 5G.

Will enterprises choose a 4G or 5G solution when looking at mobile networks for IoT connectivity?

As use of internet of things services grows, the value of connectivity services is a key concern for mobile and fixed network operators, as every sensor has to be connected, and billions of new sensors are expected to be deployed.

In the early going, no clear pattern other than “all the above” has emerged. Wi-Fi, mobile networks and fixed networks all are used, according to a Vodafone report. That could change as the business grows, but so far, there is no clear pattern.  


Going forward, enterprise executives say they are looking at both 5G and low power wide area networks as possible platforms.

Of the new LP-WAN platforms, NB-IoT, running on 4G and LTE Cat-1 and Cat M-1 are getting attention, as are the specialized LP-WAN networks.


The proportion of companies using the internet of things has more than doubled, a new report by Vodafone says.  Adoption has risen from 12 percent in 2013 to 29 percent in 2017, Vodafone says.

Transport and logistics (19 percent to 27 percent) and retail ( 20 percent to 26 percent) have shown the largest year-on-year gains from 2016.



How Close is Mobile to Becoming a Full Substitute for Fixed Networks?

The latest Federal Communications Commission report on the mobile industry in the United States, based on 2015 data that is several years old, shows the need for hard thinking about communications policy.

It is not simply the state of the mobile industry, but the coming changes that will make mobile solutions competitive with fixed network solutions in many (in some cases most) instances. That, in turn, suggests a rethink of fundamental communications policy.

As the business model for fixed networks gets worse (in the consumer space), the business model for mobile--as an access solution--gets better. How much better, compared to fixed, is the real issue.

Some policy advocates reject the notion that mobile is a reasonable and customary replacement for fixed networks. Others would contest such views.

There are many issues:
  • Can mobile really be fully effective substitute for fixed network access?
  • What will happen in the 5G era, and what will that mean?
  • What policies need to be created as landline usage continues to dwindle?
  • How much flexibility should “carriers of last resort” have in supplying demand?
  • How much demand actually exists for fixed network access?
  • How soon could mobile alternatives be viable replacements for fixed access?

One question is whether consumers actually are comfortable with a mobile-only solution to their internet, voice and messaging needs. At a high level, the number of mobile accounts in service is equivalent to the U.S. population, so the notion of full mobile substitution for fixed access seems feasible, if pricing, quality of service and other typical terms of service are equivalent.

Looking only at “human” users, the number of accounts in service is about 324 million, or nearly the population of people older than perhaps 13.

The U.S. population is about 326 million. Retail accounts used by people (not including wholesale lines or connected devices) number about 322 million.




The key point is that consumers broadly have embraced mobility as a preferred solution for their communication requirements. A growing percentage also seem to prefer mobile to such a degree that 58 percent of U.S. children 18 and under live in homes that are mobile-only.

So the notion that mobile can be accepted as a full substitute for fixed network services seems reasonable enough. The political and regulatory issue is whether service providers should be free to use more-affordable, rather than less-affordable platforms, as use of the fixed networks continues to drop.

At the moment, it appears that a typical tier-one service provider--cable or telco--gets about 40 percent to 50 percent take rates (locations that buy service from the provider). By definition, that means stranded assets--facilities that generate no income--range as high as 50 percent to 60 percent. One might argue that is inherently unsustainable.

One possible objection to viewing mobile networks as a potential full substitute for fixed networks is mobile network coverage. That does not seem to be a bigger issue than for fixed network coverage. Also, the mobility market features multiple providers (just how many are sustainable long term is an issue).

The FCC report suggests coverage by three or more competitors is 97.9 percent. Yes, there still are rural areas where coverage might be provided by only a single provider. But we have solutions for such instances. High cost support or universal service funding are the traditional remedies.

One might conclude that mobile platforms actually cannot provide fixed network equivalent service in some two percent of instances. In other cases it might still be possible, given the relative cost of fixed or mobile solutions in rural areas.


One historic problem has been that mobile network internet access speeds have lagged fixed networks.

The FCC report--based on 2015 data--suggests that median downstream speeds per device range from about 10 Mbps to 14 Mbps, on average, enough to support video streaming. Keep in mind that mobile bandwidth is supplied per device, not per location. There is no inherent requirement for sharing that bandwidth across multiple devices.

Also, speeds have increased significantly since 2015, according to Ookla. In the U.S. market, downstream speeds now top 22 Mbps. In fact, speed grew 33 percent, from 2015 to 2016, and another 19 percent from 2016 to 2017, according to Ookla.

In the 5G era, median speeds will climb perhaps an order of magnitude on a median, per-device basis, and as high as a couple orders of magnitude in some locations.



The FCC’s own tests suggest that the the main supplier 4G brands--measured on a “mean” basis--operate even faster than the median speeds, from about 14 Mbps to 19.5 Mbps.

One big issue might concern availability of multiple competitors in low-income areas. That might be a bigger issue in rural areas, compared to urban areas, simply because the cost of mobile infrastructure in urban areas is much lower than in rural areas. According to the FCC data, the number of providers of 4G service, across all income ranges, is at least four in each area.

Such data does not speak directly to the issue of affordability, but one point is that multiple providers operate in areas inhabited by people of all income brackets. In fact, availability to people in the lowest income bracket is higher than in any other income bracket.


So one possible conclusion from the latest FCC report is that the fundamental regulatory policies governing fixed and mobile communications solutions might be in need of a fresh review. “How” service is supplied might need to be reviewed, as it might not make such a difference whether hybrid fiber coax, fiber or copper access platforms, mobile or fixed wireless platforms are used.

Beyond that, there now are strong arguments to be made that former incumbents have lost the ability to shape demand, prices, terms of service and value in communications markets. In some instances, cable TV operators are the leaders. In all cases, mobile now represents a growing and preferred way to use communications.

That suggests the growing possibility that old rules which made sense when there was a true dominant carrier might need to be revisited.

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...