More Data Will be Created by Machines than People

It sometimes is hard to tell the difference between devices used by people (smartphones and other consumer devices) and those used by machines (Internet of Things). Health monitors and smart watches come to mind. They are worn by people, but “used” by servers.

Some refer to health monitor as “consume” IoT, in contrast to “industrial” IoT. The point is that there will be use cases that are hard to classify. “Smart clothing” might pose the same definitional issues.

Several years ago, for example, it would not have been unusual to find tablets classified in “connected devices” category that in some ways was the forerunner of today’s machine-to-machine or Internet of Things devices.

Perhaps few these days would count tablet connections in the IoT category. But there still are many consumer-focused appliances and machines that will be somewhat difficult to classify. Smart kitchen appliances might fall into that category.

Still, some would say connected PCs were the forerunners of today’s developing IoT markets. About the only widely-used device never really considered “IoT” are smartphones.

The point is that there is a difference between data or information created by people or machines.

source: bayforce

Perhaps nothing is clearer than the expected benefits from deploying IoT in any setting. For product and service providers, revenue upside is the expected driver of behavior, often in indirect ways such as creating better user and customer experiences.

source: Business Insider

AT&T to Launch LTE-M IoT Trial

source: Qualcomm

AT&T plans to pilot an LTE-M network in the San Francisco market starting in November 2016, followed by a full commercial launch in 2017.

LTE-M is a subset of the Long Term Evolution 4G network standard optimized for Internet of Things sensors requiring transmission speeds no greater than about 1 Mbps, as well as up-to-10-year battery life and ability to work underground.

source: Qualcomm

LTE-M technology is expected to connect a wide variety of IoT systems supporting smart utility meters, asset monitoring, vending machines, alarm systems, fleet, heavy equipment, mobile health and wearables.

Participants in the pilot include:

• Badger Meter – analyze how the LTE-M network, which is dedicated to supporting the IoT, may be used to enhance communications for smart water devices.
• CalAmp – explore how the LTE-M network can help companies more efficiently manage their connected vehicles and assets.
• Capstone Metering – demonstrate how LTE-M can improve Smart Cities sensor technologies. It will look to increase battery life and improve connectivity and sensor monitoring for underground smart water meters.
• PepsiCo – examine and test ways that sensors can improve the in-store experience with smart vending solutions for the thousands of PepsiCo products consumers love and enjoy.
• Samsung – evaluate an LTE-M-based solution to enhance performance for consumer solutions. This may include wearables or other consumer devices.

When is a Terabyte Household Data Consumption Limit a Problem?

When is a terrabyte of usage on a single consumer Internet access account a problem? When a consumer user is part of the “one percent.” That is one percent in terms of data consumption on a Comcast network in a month’s time.

Roughly, that corresponds (Comcast’s estimates) to a household consuming about 21.7 hours of high-definition format video entertainment every day of the month, based on a terabyte supporting between 600 and 700 hours of HD video, and using 650 as the median case.

Netflix estimates an hour of its HD video consumes about 3 GB per hour, though. In 2014, according to Sandvine, a cord cutter household consumed about 212 GB a month (video and all other uses).

As a rule of thumb, a typical household using Netflix and streaming video should not experience any data consumption limit issues if a 500-gbyte cap is in place, according to WhistleOut.

India Spectrum Auction Nets about 11% of Government-Forecast Revenue

India’s big spectrum auction of 2,300 MHz worth of spectrum has ended, with spectrum sold at about 11 percent of what the government projected would be the case, or roughly US$9.8 billion (if I have converted the crore properly). The government had projected sales in the $83 billion range.

As mobile executives had warned, prices for 700 MHz spectrum were simply wildly overpriced. They behaved as they spoke: nobody made a bid for any of the 700-MHz assets. Mobile executives had suggested the government lower the prices and wait before auctioning the 700-MHz assets.

Of the total of 2,300 MHz of assets, the government sold 964.8 MHz of spectrum. Mobile operators purchased about 34 percent of spectrum in the 800-MHz band, about 75 percent in the 1800-MHz band, all of the spectrum available in the 2300 MHz band and about 60 percent of spectrum in the 2500 MHz band. About 20 percent of spectrum in the 2100 MHz band was bought.

Vodafone India and Bharti Airtel were the biggest buyers of 4G spectrum, followed by newcomer Reliance Jio Infocomm and Idea Cellular.

Vodafone spent over Rs 20,000 crore, Airtel Rs 14,244 crore, Jio Rs 13,672 crore and Idea Rs 12,798 crore.

The auction results, and the squabbling leading up the auction, illustrate several important facts about the Internet ecosystem. From a mobile operator’s perspective, though spectrum access is a necessary precondition for being in business, operators cannot pay “any amount” for that access.

And mobile operators demonstrated with their wallets that spectrum prices set by the government were too high. There is experience behind that thinking. In the past, mobile operators have overpaid for 3G spectrum, for example, in India and elsewhere.

Operators have learned, from experience, that the cost of spectrum has to be weighed in view of expected revenues that can be generated by those assets.

There also are a few larger points.

Since, in the end, consumers or advertisers are the ultimate sources of all ecosystem revenue, all costs--anywhere in the ecosystem--must be matched by revenues from those sources.

The Indian auction shows that government officials and mobile operators have vastly-different expectations about the revenues that can be generated by using mobile spectrum.

There are reasons mobile operators and others might rationally expect spectrum to prices to begin dropping. For starters, much more spectrum will be made available as 5G standards are set and regulators start to release brand new spectrum in the millimeter regions.

The role of unlicensed spectrum also is growing, reducing, to a real extent, the need to buy licensed spectrum.

In some markets, spectrum sharing also will add even more resources. Finally, small cell architectures are allowing service providers to make better use of any amount of finite spectrum.

Can Internet Access be "Future-Proof?"

Remember when fiber-to-home was touted as “future-proof?” That statement is only partially true, as it was even when originally argued. The “future-proofing” is correct to the extent that physical media does not have to be replaced when a bandwidth upgrade happens (at least for perhaps 15 to 20 year lifetime of the cables).

But continual speed upgrades are part of the market dynamic. Yesterday's ceilings become today's floors. In that sense, no network is fully future proof.

To a large extent, the United Kingdom has succeeded in upgrading most locations to superfast speeds of 24 Mbps. In fact, by some estimates, about 91 percent of U.K. locations can buy Internet access at 30 Mbps. The extent of superfast coverage was about 83 percent in 2015.

Now the near-term goal is to connect 97 percent of U.K. locations at at least 24 Mbps by 2019. Back in 2009, when the idea of upgrades to “superfast” speeds became part of the national agenda, the typical U.K. consumer got speeds of perhaps 4 Mbps.

So a boost of about an order of magnitude in less than a decade is significant, if not unusual. An increase of another order of magnitude is coming, at least in part because other suppliers, using their own facilities, have leapt ahead.

Perhaps instructively, in the U.K. market, cable TV operators supply 56 percent of all “superfast” Internet access connections, though having only about 19 percent of connections.  

In late 2014, for example, the top speed supplied by BT was 65 Mbps. The top speed supplied by Virgin Media was 159 Mbps.

Since virtually all fixed network Internet access connections now are broadband, the ongoing issue is how to define “broadband.”

Source: Ofcom

AT&T to Compete with Verizon, CenturyLink Local Access Networks

AT&T now says it is deploying fixed wireless using millimeter wave frequencies to apartment complexes in Minneapolis, outside its traditional 21-state wireline service area.

In case you miss the implications, this is the first time AT&T is going to compete head-to-head with CenturyLink in the consumer local access business, in CenturyLink’s footprint, aiming to supply 100 Mbps access service to each unit in a building. AT&T says it already plans to boost speeds to 500 Mbps to each living unit.

Up to this point, AT&T's consumer operations in the fixed network area have been confined to the 21-state region where AT&T has had operations growing out of the old Regional Bell Operating Company territories.

The move is akin to Comcast announcing it is going to serve customers in a Charter Communications franchise area.

While AT&T competes directly with Verizon in the mobile business, and with both Verizon and CenturyLink in the enterprise accounts business, AT&T has not overbuilt another telco in the consumer business.

There are many reasons for that situation. For one thing, AT&T wants to avoid running afoul of informal antitrust guidelines that tend to be triggered whenever a fixed network provider serves 30 percent of available U.S. homes.

By competing out of region as a CLEC, AT&T avoids increasing the number of U.S. homes passed by its incumbent provider fixed networks.

“If successful, this will give us the ability to offer a combination of Internet, DirecTV and wireless services to apartment complexes and multifamily communities in additional metro areas.” said Ed Balcerzak, AT&T SVP.

Additional areas under consideration where AT&T might do the same include Boston, New Jersey, New York City, Philadelphia and and Washington D.C., all in the Verizon Communications footprint.

AT&T says it also is looking at Denver  Phoenix and Seattle, in the CenturyLink region.

All those efforts would have AT&T operating as  a competitive local exchange carrier competing with Verizon and CenturyLink for the first time.

"Smart Cities" Benefits Likely Will be Smaller than Projected

Asset-light business models such as Uber and Lyft are about monetizing dark vehicle assets. Airbnb perhaps is about monetizing dark room and lodging assets, also using an asset-light approach.

Wi-Fi often is an asset-light approach to mobile device access. Netlfix might be considered an asset light approach to video entertainment, at least in terms of access assets.

In other cases, big data and Internet of Things networks aim to enable more efficient use of in-use assets.

Arguably, the most-powerful trends happen when multiple values can be realized, such as combining dark assets with asset-light business models with peer-to-peer transactions and “leasing rather than owning” consumption patterns.

All those potential changes in business models should eventually affect prospects for many proposed Internet of Things services, such as “smart parking.” If vehicle ownership declines as much as some expect, there will be less demand for urban area parking, and therefore less value and demand for smart parking services.

In other words, all currently-projected markets essentially extrapolate from existing conditions. But those conditions will change as IoT and IoT-assisted ecosystems change.

Similarly, smarter transportation systems that allow users to evaluate transportation options in real time will reduce the amount of vehicle congestion the smart systems aim to solve.

Where it comes to the impact of IoT systems, feedback loops will operate, changing the context even as the systems come online. In other words, non-linearity will be a key aspect of future IoT systems. In the process of solving specific problems, the magnitude of the actual problems will diminish.

That likely will mean the expected benefits will be smaller than forecast.