Cloud Services Still are "Winner Take All"

source: Canalys

New data from Synergy Research Group shows that hyperscale operators are aggressively growing their share of key cloud service markets, a finding that should not come as a surprise. As with many application and access markets, “winner take all” of the “rule of three” tend to operate, with the consequence that only a few will lead in terms of market share.

Synergy' says 24 “hyperscale” companies in 2016 those companies accounted for 68 percent of the cloud infrastructure services market (IaaS, PaaS, private hosted cloud services) and 59 percent of the SaaS market.

Even that obscures the concentrated nature of leadership. In early 2016, for example, Amazon Web Services alone had about 33 percent share, while AWS plus Microsoft held half the market share. Add Google and IBM and four firms held about 66 percent share.

In 2012 those hyperscale operators accounted for just 47 percent share of each of those markets, says Synergy.

source: Synergy Research

In aggregate those 24 hyperscale operators now have almost 320 large data centers in their networks. The companies with the broadest data center footprint are the leading cloud providers – Amazon, Microsoft and IBM.

Each has 45 or more data center locations with at least two in each of the four regions (North America, APAC, EMEA and Latin America).

Australia Gets 4th Facilities-Based Mobile Operation

Where it is possible--where the business model works--many would agree that facilities-based competition provides faster innovation and higher consumer benefits than non-facilities-based competition.

Such facilities-based competition has been relatively rare in the fixed networks business, and arguably has gotten more difficult, while common in the mobile business. That arguably is mostly because of the high cost of fixed networks, but also because network costs are lower, and consumer demand has been higher, in the mobile segment.

And though the outcome is not yet clear, TPG Telecom has acquired 700-MHz spectrum to build its own 4G network in Australia, joining Singapore Telecommunications, Telstra Corp and Vodafone Group as facilities-based providers of mobile service in Australia.

At least for the moment, that means Australia will be among the many nations where policymakers (if not service providers) believe a four-provider mobile market is sustainable and provides the highest degree of consumer outcomes.

TPG had been the only major Australian mobile provider without its own network. TPG won two 10-MHz spectrum bands with a A$1.26 billion ($944 million) bid. It said it would spend another A$600 million building a 4G network to reach 80 percent of the population.

TPG’s ambitions are quite high, as it presently has market share in the range of just three percent.

source: mobilenetworkguide.com

%Share	Total Market June 2016
Telstra	41.8
Optus	21.8
Vodafone	15.2
Virgin	5.4
Boost	0.7
amaysim/Vaya	4.5
ALDI	2.2
TPG/iiNet	2.8
Other MVNOS	5.5

source: Kantar Worldpanel

Virtualized Access a Growing Foundation for Mobile

In both the core and access portions of the communications network, virtualization is becoming a fundamental building block for use of all network assets, owned and borrowed. Such sharing has been informal for some time, as exemplified by smartphone Wi-Fi access. But sharing mechanisms are becoming more formalized and integrated, as exemplified by

Wi-Fi calling.

In many cases, Wi-Fi calling is a way to overcome mobile network signal issues indoors, with sessions maintained even when a connection switches from mobile to Wi-Fi, for example.  

There are, in fact, a growing number of ways to  virtualize spectrum and access assets, both formal and informal. Several methods for bonding licensed and unlicensed spectrum assets are available, including LTE-U, license assisted access and license shared access.

In a growing number of instances, both direct and indirect revenue models might be involved, with a growing amount of integration of owned and third party access assets.

Of course, in other cases, Wi-Fi is a feature of a service the buyer actually does pay for directly. The best example is consumer fixed network internet access, where Wi-Fi typically is the physical mechanism used to connect devices. In that sense, Wi-Fi replaces physical cables and outlets.

Wi-Fi calling also represents an early and informal method of spectrum sharing, in this case treating licensed mobile operator assets and owned or third-party Wi-Fi assets as interchangeable parts of the physical network access infrastructure.

That virtualization of the details of network access already is a building block for service providers

Such as Republic Wireless and Google's Project Fi, and is different from Wi-Fi-only calling offered by over-the-top apps such as Skype, Google Hangouts, Facebook Messenger and WhatsApp.

In the future, MulteFire, a method for creating Long Term Evolution 4G networks entirely using unlicensed spectrum, might be an important platform.

Consumers Still Do Not Have a Good Grasp of Value, Price in Internet Access

It never has been easy for consumers to figure out “how fast, or how good” their internet access services are, how much they consume, or how much that consumption actually costs. Matters are complicated by the fact that value keeps climbing, while costs per unit keep dropping.

In 2009, the median consumer internet access speed was just 7 Mbps, according to the Federal Communications Commission. But by 2016, speeds had increased by an order of magnitude or even two orders of magnitude, among the tier-one providers.

And even if prices dropped, consumers have changed behavior, upgrading to faster tiers that cost more.

source: Ookla

All that matters, as a recent survey of consumer attitudes suggests most U.S. residents support building of municipal broadband networks when service is deemed inadequate.

A substantial majority of the public (70 percent) believes local governments should be able to build their own broadband networks if existing services in the area are either too expensive or not good enough.

Of course, there also is evidence that U.S. residents are ill-equipped to make such determinations.

When asked what download speed they have for their own home internet service, 47 percent were not able to provide an answer. Though 73 percent of U.S. homes have fixed network access, and probably 13 percent are mobile-only for internet access (implying that possibly up to 86 percent of the population buys internet access), about 25 percent of respondents guessed that less than half of U.S. residents had fixed network internet access.

On the matter of retail cost, everyone agrees that cost per gigabyte (consumption), or cost per Mbps (speed) have declined over time, and continue to do so. Beyond that, effective cost per gigabyte depends largely on how much any particular user consumes, compared to the posted rates. In other words, if a consumer buys 10 Gbytes of usage per month for $30, but uses only one GB, then the effective cost is $30 per used GB, even if the nominal cost is $3 per GB.

Volume also makes a huge difference: buy more and price per unit drops sharply.  

Typically, retail pricing in the U.S. mobile market is about $10 per GB or less for plans with moderate usage buckets, though pricing for small buckets can reach $15 per GB.

The new unlimited plans make such determinations even more problematic. But it would not be unusual for one GB to cost between $3 and $4 per GB, when a plan has some volume.

Fixed network prices tend to be lower than that, often by an order of magnitude (10 times). The point is that U.S. consumers cannot easily figure out whether value and prices are in line, appropriate or reasonable.

source" The Next Platform

FCC Does Not Plan to Allow in-Flight Calling

The Federal Communications Commission looks to continue its present policy of no in-flight phone calls. Chairman Ajit Pai said  “I stand with airline pilots, flight attendants, and America’s flying public against the FCC’s ill-conceived 2013 plan to allow people to make cellphone calls on planes.”

“I do not believe that moving forward with this plan is in the public interest,” Pai said. Pai plans to take that policy off the agenda, terminating the FCC’s 2013 proceeding to relax the rules and allow in-flight calling.

“Taking it off the table permanently will be a victory for Americans across the country who, like me, value a moment of quiet at 30,000 feet,” Pai said.

5G Networks Will be More Expensive

Business models in the 5G era are likely to be more complicated for a number of reasons, including the likely higher cost of infrastructure. 

For starters, millimeter wave frequencies will require use of small cells, which means many more sites and therefore lots more backhaul locations. Support for new internet of things apps might require--or benefit from--overlay networks of the low power wide area type. So one network might not “do it all.” 

 Radios are likely to feature “massive” multiple input, multiple output (MIMO) technologies, which are more complex than today’s mobile radios, and incorporate new beam-forming technology as well. 

 Also, in addition to designing new mobile networks incorporating millimeter wave frequencies, at least some access providers might choose to build fixed wireless networks that coexist with millimeter-wave mobile networks. 

That will require more frequency coordination. On the revenue side, new business models and roles in the internet of things and machine-to-machine communications areas will have to be created, as services sold only to human beings might not represent all that much incremental revenue.

Most Believe There is a Business Case for Millimeter Wave Networks

There might still be some observers who think millimeter wave networks will “not work,” in the sense of a business case that allows 5G and other wireless access providers to make money.

Few matters are as contentious as the business model for millimeter wave communication, either in a mobile or fixed wireless deployment. The biggest concerns probably concern signal propagation, and the impact on business models.

Some think sheer reach will be an issue, while others might argue that the cost of infrastructure, relative to revenue, will prove sub-optimal. Others, privately, simply think propagation or deployment issues will prove insurmountable.

The Mobile Experts model for 5G deployment estimates the cost of a 5G network and the potential revenue from new 5G services such as fixed wireless might be quite challenging. “There is simply not enough high-density demand for that,” Mobile Experts says.

To be sure, density is a key issue, as high desnity of users is needed to make a business case for small cells. But that might be a feature, not a bug.

So far, most observers and most mobile operators appear to believe small cell networks using millimeter waves will work, and will have a business case.

AT&T is acquiring Straight Path Communications, a holder of spectrum licenses in the 28 GHz and 39 GHz region. The deal means AT&T will acquire 735 licenses in the 39 GHz band and 133 licenses in the 28 GHz band. These licenses cover the entire United States, including all of the top 40 markets, to support the coming 5G network.

That move is but the latest in a series of developments indicating that millimeter wave spectrum and small cells will be foundational for coming 5G networks. Previously, AT&T had acquired 24 GHz and 39 GHz bands covering 8.4 billion MHz POPs as part of AT&T’s acquisition of FiberNet.

Dish Network, for its part, acquired from sister company EchoStar spectrum licenses covering four markets in the 28 GHz band. THat move adds more spectrum to the trove Dish Network has amassed, either to build its own mobile or wireless network, or as an asset to be sold.

Verizon, for its part, as part of its acquisition of XO Communications, gained

102 licenses in the 28 GHz and 39 GHz bands, covering 188 billion MHz­-POPs, or 23 times the FiberTower assets.

Separately, Charter Communications has applied to the Federal Communications Commission to test 28 GHz frequencies.