Nobody Knows What the Cost Structure of a Modern Network Will be in 10 Years

You cannot assume the cost of building a high-performance Internet access network is going to be the same, in 10 years, as it is today. How much different is the only issue. Some trends are easy to identify.

The cost of supplying a gigabit connection on a 5G network likely will be lower than on a fixed fiber to the home network, and possibly lower than a cable TV hybrid fiber coax connection.

But that is not all that will change. Google Project Loon is testing entirely new platforms for Internet access, using fleets of balloons. Both Google and Facebook are testing use of unmanned aerial vehicles. New fleets of low earth orbit satellites will change the cost of delivering Internet data by satellite.

Fixed wireless technology being developed by Facebook, Google and others such as Starry  likely will change that cost curve as well.

And that is not all. An effort to develop standards-based and open telecom platforms is underway by the Facebook-initiated Telecom Infra Project that has gained important support from mobile and fixed telcos and many equipment suppliers.

Followng on the heels of Facebook’s Open Compute Project (OCP) to create open standards for data centers, the Telecom Infra Project now seeks similar results for access platforms.

“A few years ago, Facebook was faced with a data center problem familiar to many scale companies: We depended on proprietary systems and hardware that were inflexible and expensive,” said Jay Parikh, Facebook Global Head of Engineering and Infrastructure. “We realized quickly that this approach would not be sustainable; we needed to find a new way.”

Note the language: traditional rack and stack approaches were “unsustainable.”

The end result, for Facebook, was that “we were able to...save billions of dollars in infrastructure costs over the last few years,” Parikh said. The obvious winner was Facebook and its users. The obvious losers were suppliers of traditional data center gear.

“We recognized that telecom infrastructure could benefit from the same innovations taking place in the data center,” Parikh said. So make note: the winners will be Internet access providers. The losers will include many suppliers of network platforms, or whole lines of equipment and software platforms.

“It was clear that the raw building blocks of what we were developing for our own infrastructure could be applied to telecom networks with great benefit,” he said.

The Telecom Infra Project “ is bringing together operators, infrastructure providers, system integrators, and other industry players to work together to develop new technologies and rethink approaches to deploying network architecture.”

Early founding members include Intel and Nokia, Deutsche Telekom and SK Telecom.

At first, “TIP will focus on disaggregating the components of network infrastructure that are traditionally bundled together and vendor-specific,” said Parikh.

TIP members will work across three areas: access, backhaul, and core network management.

As one early example, Facebook has been working in partnership with Globe, deploying a low-cost, solar-powered network-in-a-box solution, bringing mobile coverage to a village. “In the first week alone, we connected more than 60 percent of the community,” said Parikh.

New members include Axiata Digital, Indosat, MTN Group, Telefonica, Vodafone, Acacia, ADVA, BlueStream, Broadcom, Coriant, Deloitte, Juniper Networks, and Lumentum.

The TIP Board of Directors includes Dr. Alex Choi of SK Telecom (TIP Chairman), Axel Clauberg of Deutsche Telekom AG, Ashish Kelkar of Facebook (TIP Secretary and Treasurer), Lynn Comp of Intel, and Henri Tervonen of Nokia.

Project groups also have been created to address “the most pressing industry needs including connecting the unconnected or underserved populations, and augmenting the development of powerful new technologies like 5G.”

The access system integration and site optimization group is chaired by SK Telecom

The unbundled solutions group is co-chaired by SK Telecom and Nokia, and will seek cost-effective, low-power and low-maintenance solutions.

Media-friendly solutions, chaired by Intel, will focus on mobile experience, especially for close-to-edge solutions.

In the backhaul area, Facebook heads the effort to develop “thin and extensible software stack to autonomously coordinate routing, addressing and security related functions in packet-switched IPv6 networks.”

The open optical packet transport project is co-chaired by Facebook and Equinix, and is working on Dense Wavelength Division Multiplexing (DWDM) open packet transport architectures that avoid supplier lock-in.

The core network optimization project is chaired by Intel, and seeks to disaggregate

core network components.

The greenfield telecom networks group is co-chaired by Nokia, Facebook and Deutsche Telekom, and will work on IT-based network architecture.

If and when those solutions emerge as commercial realities, we must assume the cost structure of networks will be lower. So all current assumptions about business models will have to be revised as well.

Huge Shift in Local Government Thinking and Policy Leads to Gigabit Internet Access Investment

A really-major shift in thinking by municipal officials has occurred, where it comes to permitting, licensing or franchising high speed Internet access businesses.

In the past, it would have been impossible for any telco or cable TV company to get permission to build a network unless that network reached all homes in a city or town.

The new thinking--spurred by Google Fiber--is that Internet service providers should be allowed and encouraged to build gigabit networks wherever there is demand within a city, without requiring ubiquitous network builds.

That is the reason Verizon Communications now is building out gigabit connections in Boston, for example, when the original business case--assuming a ubiquitous build--was not deemed attractive.

"The past administration here wanted the sort of buildout we have done in other areas where you build everywhere and you go in and get penetration," Verizon CEO Lowell McAdam said. The new administration, on the other hand, is "more willing to help us get rights of way, help us push fiber into the neighborhoods, and do more pre-subscription a la the Google model."

In other words, by allowing ISPs to build only where there is a reasonable expectation of demand, high-performance networks are being deployed where they would not have been built before.

In part, that change followed the example of the competitive local exchange carrier business, which allows service providers to build facilities only in areas where there is business customer demand for high-bandwidth services.

That, in turn, was not lawful until passage of the Telecommunications Act of 1996. The new attitude by local government officials is leading to more intense deployment of gigabit facilities than would be the case if all builds had to be ubiquitous across a city.

That is a really big deal.

SE Asia is World's Fastest-Growing "Internet" Market

With 260 million users now and an estimated 480 million by 2020, Southeast Asia’s is the world’s fastest growing Internet market, adding 3.8 million new users a month, according to an analysis by Temasek, sponsored by Google.

E-commerce in the region could be worth $88 billion, growing at 32 percent per year, while

The online travel market could reach $90 billion, growing at 15 percent per year.

Online media such as games and advertising could grow to $20 billion, increasing at 18 percent per year, while the taxi on-demand service could hit $13 billion and 29 million monthly riders, up from an estimated 7.3 million at present, growing at 18 percent per year, the analysis suggests.

There are obstacles to overcome, ranging from a relative lack of human resources (engineering and managerial), access to capital and consumer mistrust about the security of online transactions.

But most of the obstacles are directly related to access and other forms of infrastructure. Logistics networks required for remote product delivery are underdeveloped. Some 60 percent to 70 percent of consumers are outside the traditional banking system.

Low Internet usage, especially for archipelago nations such as Indonesia and the Philippines, also are key obstacles to be overcome.

Smartphone-Only Internet Access in U.K. Grows to 16%, Up 10% in One Year

A study sponsored by Ofcom might suggest that 16 percent of U.K. Internet users rely on devices such as smartphones and tablets for online access, and do not primarily use fixed Internet connections.

In 2014, mobile-only access was at about six percent. Ofcom therefore suggests that use of computers for online access has decreased by 10 percentage points--from 81 percent to 71 percent--since 2014.

One might argue that a qualitative survey also sponsored by Ofcom shows the complexity and nuances of smartphone-based Internet access that could have implications for developing regions as well.

“The extent to which smartphones are liberating or limiting for these participants is complex, nuanced and highly dependent on their circumstances,” Ofcom says. “In addition, the perceived price of using data lead many users to limit usage.”

In fact, “heavy reliance on smartphones may inhibit users from developing alternative digital skills, such as typing,” the survey suggests. Perhaps that will not be a limitation, eventually, but Ofcom suggests lack of such skills is a job-affecting circumstance.

Also, many users who were mobile-only had more-limited skills in terms of technical troubleshooting and file and information management, Ofcom says.

Aside from other issues, the cost of mobile data was an issue. “The limitations of smartphones as a primary means of going online, across all parts of the sample, included the perceived pressure to complete tasks quickly to prevent the erosion of their data allowances,” Ofcom notes.  

“Creating, editing and sharing any document of length in office software applications (e.g. MS Office) was seen as almost impossible for most participants,” the report notes.

Those issues were largely experienced by users who were smartphone based by circumstance (low income, no at-home fixed network access, no home PC, for example), not choice (some users have jobs where they are highly mobile, so phone-based access is seen as an advantage).

The implication might be that smartphone-based access, which will be the way most people, in many regions, will use the Internet, does pose some challenges.

The cost of using mobile data is an issue. The difficulty of creating and manipulating documents and files are other issues.

Is "Zero" a Price?

Is “zero” a price? The answer matters, because there are many instances where regulators insist they are not in favor of rate regulation, at least when markets are competitive.

In other words, whenever a regulation mandates, or prohibits a “zero” price, that is an instance of rate regulation.

This matters since banning or enabling zero pricing is rate regulation, even when rate regulation is expressly not within policy guidelines. That recently has become an issue where zero rating of mobile Internet access is outlawed.

The essence of rate regulation is that some prices are not allowed, while other prices are mandated. Generally, we are accustomed to prices being limited to keep them low.

What is different with bans on zero rating is that the restriction applies in the other direction: governments make illegal the offering of “no incremental charge” or “free” access.

Generally, price regulation is used when an entity has a monopoly or a dominant market position that gives it excessive market power. So prohibiting “free” as a price implies concern about monopoly power.

Regulators sometimes insist they are not engaged in rate regulation when mandating that prices of zero are not lawful. That is wrong. Zero is a price, and mandating or prohibiting zero rating is rate regulation.

The answer also matters because rate regulation--whatever its other suggested benefits, also tends to depress investment.

That means rate regulation, all other things being equal, leads to lower investment by competitors in a market who own facilities, even if it increases demand by customers who want to use that rate-regulated network.

To the extent that bans on zero rating impose price rules, it is a form of rate regulation, even if not formally a matter of formal “rate setting.”  While paid services have prices that are acknowledged to be “rates,” it is much less obvious that “zero” also is a price.

How Will AT&T's Virtualized Network Affect its Suppliers and Retail Sales Partners?

It isn’t as easy to describe the likely impact of network functions virtualization (NFV) on the retail end of the business as it is to envision the implications for network infrastructure supplier impact.

By definition, virtualization means new and existing apps can run on a single hardware platform, rather than having dedicated equipment for each app. That should mean less spending on network equipment, so less capital expense, and arguably lower operating expense.

At a high level, that also implies lower revenue for telecom network equipment suppliers.

This will be a big deal for AT&T and many of its networking equipment suppliers.. “We are now forecasting that by the end of this year 30 percent of our network functions will be virtualized,” said Ralph de la Vega, AT&T vice chairman. “ And by 2020, it will be 75 percent.”

“we’re moving all of our service to this on-demand platform over time,” he said.

In principle, there should be implications for retail sales, or on-going customer support operations. As AT&T rolls out “Network On Demand,” retail enterprise customers will be able to  dial up, or dial down, bandwidth on demand. “If the customer has a five megabit per second circuit and they want to go a 20, they go to the portal, they make a change and in less than 90 seconds, the new service is provisioned,” said de la Vega.

“There is no need to call a sales person, there is no need to order equipment, there is no need to setup another connection it happens in 90 seconds,” de la Vega said.

For AT&T, that could often translate into higher revenue, if the direction of changes by most customers is “up,” rather than “down.”

“And what I love about it is my revenue cycle is 90 seconds,” said de la Vega.

In principle, that same bandwidth on demand feature could be available to medium-sized or smaller businesses as well.

De la Vega used the example of a small cancer center with 12 locations. That sort of mid-market customer can, assuming the bandwidth is available at all locations, dial up and then dial down bandwidth, allowing it to send big files episodically, at any of the locations.

That could have some impact on retail sales channels, if the “moves, adds and changes” revenue becomes a matter of customer self service.

That could drastically limit the amount of MAC revenue any retail agent--or AT&T itself--earns from bandwidth services.

AT&T has More than 50% Market Share of U.S. Connected Cars

AT&T Unlimited Plan customers now can add select connected cars or a ZTE Mobley vehicle Wi-Fi plug-in device to their data plans for $40 each month, getting unlimited data use for the connected vehicles.

Unlimited Plan customers also can buy 1GB of data per month for $10 per connected car or ZTE Mobley.

Connected car customers can continue to buy stand-alone data plans or add their car to a “Mobile Share Value” plan.

AT&T says it has more than eight million connected cars on its network at the end of the first quarter of 2016, representing more than 50 percent of all new connected passenger vehicles in the United States,on the strength of partnerships with 19 automakers, including Porsche, Jaguar, Land Rover, Infiniti, Audi and Volkswagen.

Embeded connectivity and and the ability to integrate smartphones ultimately are expected to drive most of the market. Simple tethering of smartphones will be a factor, however, according to analysts at Juniper Research.

source: Juniper Research

IoT Revenue Mostly Will Happen at App Layer

Platforms often drive much more economic activity than is directly generated by the platforms themselves, and that is virtually certain to the case for access and connectivity markets created by Internet of Things apps and devices.

A McKinsey Global Institute study suggests that the Internet of Things has the potential to create economic impact of $2.7 trillion to $6.2 trillion annually by 2025. Connectivity (access) might account for about 16 percent to 17 percent of total ecosystem revenues, according to McKinsey analysts.  

source: Woodside Capital

Vodafone plans to roll out the narrowband IoT (NB-IoT) platform to support Internet of Things connectivity across multiple markets in 2017, said Vodafone Internet of Things group Director, Erik Brenneis.

As often is the case when new platforms are being commercialized, different tier-one mobile operators are making different bets.

Altice’s SFR is deploying Sigfox while rivals Bouygues and Orange have opted for LoRa. Orange, though, has not ruled out use of other platforms as well.

China Mobile also has chosen to support NB-IoT.

In some ways, the choice of NB-IoT is not too surprising. Historically, tier-one telcos have preferred solutions based on licensed spectrum, and also have preferred options that build on already-established networks and also have perceived “quality of service” advantages. BB-IoT fits within that approach.

“NB-IoT operates in licensed spectrum and that is important to us at Vodafone because we need to deliver a high quality experience to our customers,” said Brenneis.

All of the proposed networks emphasize lower-cost connectivity and lower power consumption.

Facebook Explores Anchor Tenant Model in India

In the U.S. market, many innovators have looked to the concept of “anchor tenants” as one way to build a sustainable Internet access service. Schools, hospitals or government buildings often are seen as suitable anchor institutions.

Some would-be suppliers of gigabit access, for example, think it makes sense to connect anchor institutions that essentially become hubs for more-extensive networks surrounding the anchor locations.

The same concept underpins many rural Internet access efforts, which first try to connect schools, for example, and then use the schools as hubs. In some ways, that is the concept behind village kiosks that are connected by satellite links, and then use Wi-Fi for local distribution.

Facebook‘s “Express Wi-Fi” uses the same concept, encouraging anchor tenants to become local Internet service providers, providing a sustainable business model.

“We are already live in India and Indonesia with Express Wi-Fi,” said Ryan Wallace, Facebook technical program manager. “If you look at Myanmar as a great example that has got poor fiber infrastructure so satellite would probably be an interesting play there.”

“The Philippines is an island nation, so fiber is very much within the metropolitan areas. Satellite could be a big play there.”

Facebook has partnered with an Indian rural Internet access provider, AirJaldi, to manage the actual installation and operation of the Express Wi-Fi service.

For AirJaldi, which is based in Dharamsala and mostly provides Internet services to large customers like the Dalai Lama’s Tibetan government in exile, the project is a chance to figure out how to make money serving a widely dispersed customer base.

Google Play Android Apps to Run on Chromebooks

In a possibly-huge move, Google has announced that all Google Play apps will run on Chromebooks. As a practical matter, for Chromebook users, that means a Chromebook will be able to run Skype and GoToMeeting, for example, apps that previously required a download, and therefore could not be run on a Chromebook.

Essentially, all apps that run on Android phones and tablets will now run on Chromebooks without compromising their speed, simplicity or security.

The advantage is about one million applications.

Up to this point, Chromebook functionality has been limited by the fact that everything had to run within a browser.

Every “app” was essentially a browser plug-in. High security and fast startup were clear advantages.

Apparently, the Chromebook and Android teams to take advantage of innovations in enabling the Android environment work within the ChromeOS “container.”

Since Chromebooks are running desktop class hardware, the Android apps consume virtually no overhead, the user experience is consistent, and the environment is secure.

The Google teams say the Chrome operating system is not rendering the Android environment, running the apps fully.

The feature will start rolling out in the developer channel with M53 on the ASUS Chromebook Flip, the Acer Chromebook R 11 and the latest Chromebook Pixel.

Over time, this will roll out to other Chromebooks in the market as well.

15% of U.S. Residents Have Used Ride-Sharing, 11% Have Used a Home-Sharing Service

An estimated 15 percent of U.S. adults have used a ride-hailing app such as Uber or Lyft. Some three percent of U.S. adults use ride-hailing apps on a daily or weekly basis, and around 66 percent of these regular ride-hailing users indicate that they own a car or regularly drive a personal vehicle.

About 11 percent have used a home-sharing service such as Airbnb. U.S. residents 35 to 44 are nearly twice as likely as those ages 18 to 24 to have used home-sharing services (16 percent compared to nine percent).

A majority of regular ride-hailing users are car owners or drivers, but are significantly more likely to use a range of other transportation options.

source: Pew Research Center

Ride-hailing apps appeal heavily to younger adults. The median age of adult ride-hailing users in the United States is 33, and 18- to 29-year-olds are seven times as likely to use these services as are those age 65 and older (28 percent compared to four percent).

Ride-hailing use is also heavily concentrated among urban residents (especially younger urbanites and those with relatively high levels of income and educational attainment), while being consistently low among rural residents of all kinds.

In Some Countries, 4G is Faster than DSL

With the caveat that the situation changes quickly in the U.S. high speed access market, recent speed tests conducted by DSL Reports suggest that, in many countries, under some conditions, users might experience faster speeds using a Long Term Evolution 4G mobile connection, rather than an all-copper digital subscriber line service.

Of course, all things are not equal. DSL cost per bit is far lower than mobile data cost per bit, so the typical application set for any single user will determine whether mobile data is a functional substitute for DSL.

The DSL Reports tests suggest an average DSL speed of about 13.7 Mbps, and an average 4G speed of about 12.3 Mbps. There is not so much functional difference there, for a single user account.

The cost decision of course hinges on how much data any single user consumes in an average billing period. The economics shift dramatically for multi-user households.

The tests show that in some countries, such as Canada and Australia, 4G speeds actually are faster than DSL.

source: DSL Reports

In another year or so, cable speeds are likely to climb fast, as Comcast and other providers roll out ubiquitous gigabit access, and fiber providers upgrade many lines to 1 Gbps as well.

Out another five to 10 years, matters could change again, as mobile 5G speeds potentially reach a gigabit per device, initially, then potentially extend to 5 Gbps or 10 Gbps.

What are Possible Business Models for Managed Services, if Zero Rating is Unlawful?

The Telecom Regulatory Authority of India has banned zero rating the ability of an Internet access provider to offer access to some Internet apps without requiring purchase of a data plan, or levying usage charges for use of those apps.

But TRAI has not issued clear guidance about what service providers will be allowed to do, in terms of creating managed services. TRAI recently has suggested that no currently-offered Internet app or service can be refashioned as a managed service.

Some might argue that zero rating is precisely what will eventually be required to create incentives for the “unconnected” to experiment with and then become accustomed to use of Internet apps. The reason is simply that mobile data is too expensive for many non-users who do pay for mobile voice and text messaging services.

source: Percolate

Now TRAI appears to be attempting to create such frameworks, the key issue being permissible revenue models. So far, TRAI has suggested that some current models might make sense.

One possible  model is to offer an incentive or reward to users when customers download certain application or take some defined activity on a particular website.

In this reward based model apps may provide rewards in the form of a recharge for data usage or for voice usage to the users, TRAI suggests.

Perhaps ironically, TRAI suggests another possible model is sponsored data, where there is a “toll-free” approach. “This model is also prevalent in many developed markets, allowing free access to certain websites and applications,” a TRAI consultation paper says.

A reimbursement model might also be permissible, TRAI suggests. In that model, the user pays for usage, but then some third party reimburses the user.

Every country has a distinctive way of approaching regulation and economic freedom. India is not an exception in that regard. Though it is not the current approach, some might argue the simplest approach is simply to allow zero rating.

High Speed Access: U.S. Behind, at Par, or Ahead?

Evaluating the state of Internet access in the U.S. market always is a contentious matter. These days, it also is a rapidly-changing matter.

Some argue the U.S. high speed access market is way behind some other countries, though it might be fair to note that virtually all those other countries are small, or compact, highly urban or places where access is subsidized heavily by the government.

The point is that it always is much easier for a city-state to boost its infrastructure performance, compared to any continent-sized country, especially any country with large areas that are sparsely settled.

Also, such comparisons also can change quite rapidly. Very soon, the U.S. will be the market with the highest percentage of gigabit per second availability, based almost entirely on the decision by Comcast--the largest U.S. ISP--to boost 100 percent of its connections to gigabit speeds.

Other U.S. cable TV operators also will do so, as the incremental cost is as low as $200 per location to upgrade to gigabit speeds.

source: Point Topic

Others point to the retail price of access, claiming that prices are high in the U.S. market. On a percent of income basis, though, U.S. prices are quite low.

Also, some continue to complain that consumers do not have enough choice when it comes to broadband providers. But there is a trade-off between competition and investment, because profits are affected.

According to Akamai’s recent State of the Internet report, all 50 states, plus Washington, D.C., saw increases on a year-over-year basis in average connection speeds, and those ranked among the top 10 experienced double-digit gains, says the Information Technology and Innovation Foundation.

“The fact that the U.S. broadband industry has achieved competitive speeds while also maintaining low entry-level pricing is remarkable considering the hurdles we face with sprawling suburbs, rural states, relatively low levels of computer ownership, and relatively high rates of poverty,” says ITIF.

But competition in broadband networks is not an unalloyed good—more competitors in a given geographic market is not always better, ITIF also says.

Having more firms with smaller market shares competing at low margins will necessarily raise overall production costs while reducing average firm revenues.

The result, therefore, will be higher prices overall.

As such, spurring more competition through proactive government subsidies or other policies is almost always less efficient in lowering prices and improving service than effective competition among fewer firms, the foundation argues.

Global Business IP Telephony Market $35 Billion by 2018?

The global business VoIP market is projected to grow to $35 billion in service revenue by 2018, up from from $24 billion in 2013, according to Infonetics Research.

The hosted unified communications business will generate about $4 billion in annual revenue, while hosted PBX services might generate about $8 billion in annual revenue.

source: Infonetics Research

Voice over IP (VoIP) and unified communications (UC) services are forecasted to grow from $68 billion to $88 billion by 2018, according to Infonetics Research. These figures include both residential and business sectors as well as access services (SIP trunking) directly related to use of voice and unified communications services.

Other forecasts are not quite so robust. Frost and Sullivan has estimated that 2018 IP telephony and UC services might reach about $8 billion by 2018.

source: Frost and Sullivan