Cost of Using Internet Access Drops, Globally

By 2025, entry-level (fixed network) broadband services should be made affordable in developing countries at less than two percent of monthly gross national income per person. That matters as the cost of using internet access services as a percentage of income is a key measure of affordability.

More importantly, the total number of active mobile broadband subscriptions is expected to reach 4.4 billion by end 2018, up from 3.3 billion, at the end 2015,  the International Telecommunications Union says. That matters since mobile internet access is the way most people in developing countries use internet access services.

This is a clear case of perceiving a “glass half empty, or half full.”

In January 2017, the Broadband Commission lowered the de-facto standard for Internet  affordability to two percent of average income, from the previous five percent levels, evidence of significant price declines.

Although the majority of the world’s population (52 percent or 3.7 billion) currently remain unconnected, 3.8 billion people or 49 percent of the global population will be online by the end of 2018.

In less developed countries, prices fell from 32.4 percent to 14.1 percent of GNI.

The point is that, when making cross-country comparisons, costs must be adjusted for purchasing power.

Around 1995, the cost of buying a U.S. business connection supporting a kilobit per second might have been US$1.50 to $1.75. In other words, a 56 kbps connection might have cost as much as $98 a month.

By about 2006, even consumer internet access costs had dropped to about two cents per kbps. So a 10 Mbps connection might then have cost the same as the 56 kbps connection of 1995. In 2017, U.S. 100 Mbps connections cost about the same as a 56 kbps connection of 1995.

As speed has grown and apps have evolved, consumers now use more data (megabytes), so the cost per consumed megabyte also has fallen, even as people use more data.

Source: Boston Consulting Group

While complaints about high prices never seem to stop, in developed markets as well as the United States, the percentage of disposable income spent on fixed network internet  access is about 1.7 percent of gross national income per person.

The glass is half full.

"Factual" and "True" Observations about Internet Access Quality

Some statements are factual, but arguably not “true.” It is factual that fixed network or terrestrial network coverage gaps exist in rural and other “hard to reach” areas. In many rural areas, especially mountainous areas where few people live, there might be zero mobile network coverage, to say nothing of fixed network coverage.

The existence of such gaps might, or might not, bear much relationship to the state of service quality in dense, suburban and other areas with greater population density. In other words, it is not a “failure” of government or industry that some areas have poor to no terrestrial network coverage. Some areas simply have such low population density that only satellite service is commercially viable, even with deployment subsidies.

The simple reality is that coverage of the “last couple of percent” of people in most countries with rural, mountainous or island geographies is quite expensive. In the U.S. market, for example, it is coverage of the last one percent of people that is most tellingly expensive.

The point is that coverage gaps can exist without necessarily telling us very much about the state of internet access on a wider basis within any country.

source: ITU

source: FCC

5G Millimeter Wave Capacity: Bits per Hertz Matters

There are several reasons why the advent of millimeter wave spectrum for 5G vastly increases bandwidth, and thereby creates new business opportunities for mobile operators. Not only will millimeter wave spectrum represent a vast increase in mobile capacity (an order of magnitude to two orders of magnitude effective new spectrum), but millimeter wave spectrum also is more efficient.

Where spectrum below about 2 GHz has a spectral efficiency up to 2.5 bits per Hertz in a 4G context, and up to 3.8 bits per Hertz on a 5G network, millimeter wave spectrum has an efficiency up to seven bits per Hertz.

Basically, not only does millimeter wave spectrum represent an order of magnitude more capacity (Hertz), it also represents more bits per Hertz, as much as double what is possible on a 5G network using spectrum below 2 GHz or so. The reason has much to do with frequency and its relationship to symbol representation.  

source: T-Mobile US

By 2028, 90% of All 5G Traffic Will be Video

The importance of content, especially video content, delivered on mobile networks can be glimpsed from a new forecast by Ovum. By 2028, about 90 percent of 5G traffic is expected to be video.

Between 2019 and 2028, Ovum analysts predict, media and entertainment companies will compete for about $3 trillion in cumulative mobile content revenues, of which about $1.3 trillion will be earned on 5G networks, Ovum suggests.

By 2025, 57 percent of mobile revenue globally will be earned on 5G networks, say researchers at Ovum. By 2028, Intel and Ovum expect that number to rise to 80 percent.

he report, sponsored by Intel,  predicts that augmented reality and virtual reality will generate cumulative revenues of $140 billion (£106 billion) between 2021 and 2028.

Immersive and new media applications which don’t even exist today could generate $67 billion (£50.8 billion) a year by 2028, equivalent to the value of the entire global media market in 2017, including games, music and films, the study suggests.

Colt Technology to Launch Virtual Network in 2019

Colt Technology Services Group says it plans to start a three-stage company-wide deployment of NFV capabilities in 2019. What will that entail? The ability to use generic universal CPE (uCPE) supporting virtual firewalls, cloud-based WAN acceleration and SD-WAN.

To complicate matters, the use of generic CPE might clearly be an instance of NFV, but virtual firewalls, WAN acceleration and SD-WAN might properly be considered SDN applications.

That illustrates neatly the problem we have when describing network virtualization.

As a practical matter, it sometimes can be difficult to understand precisely what a “virtual communications network” actually does. It also can be difficult to understand how a "virtual" network is created, as that most often includes a mix of changes broadly including both network functions virtualization and software defined network adaptations.

Network functions virtualization (NFV) is one key aspect of virtualization, but not the only key aspect. It also is hard to understand what a “network function” is, in terms of “virtualizing” it.

source: fiber-optic-equipment.com

Software defined networking is the other key building block, and arguably is even more important where it comes to new service creation, where NFV mostly is about lower capital investment and operating cost.

So what are examples of network functions that can be virtualized? Routers, mobility management,policy and charging rules, session border controllers,session initiation protocol and media gateways. Providing IP Multimedia Subsystems functions (IMS) provides another example.

source: Wikibon

That is the terrain of NFV: separating data and control planes and allowing compute and control functions to be moved back from remote network elements to more-centralized locations.

SDN is more centrally related to enabling remote control of services. Software-defined wide area networks (SD-WANs) provide one good example. But supplying firewall, antivirus, video or parental control services are examples of SDN virtualization.

Note that, in general, NFV deals with network optimization, while SDN tends to involve customer-facing features.

source: Spirent

source: SDX Central

Nokia Launches Fixed Network "Network Slicing"

In a move with huge potential implications, Nokia has launched its fixed access network slicing solution, allowing fixed network service providers to create virtual networks as mobile operators will be able to do on their 5G platforms.

That might potentially enable full control of virtual networks that allow many new providers (app providers, platform providers, device suppliers, mobile virtual network operators, content providers) to essentially create their own national or global networks quickly and flexibly, with differentiated network features, to an extent.

Sure, entities have been able to construct private networks using traditional wholesale purchase agreements. But network slicing should allow faster, easier, more flexible flow-through networks all the way to the network edge. Over time, such virtual networks also should be less costly.

Network slicing allows fixed network operators to “scale to a virtually unlimited number of discrete network slices that can be independently operated, for example to run 5G mobile transport, wholesale or business services,” says Nokia.

A network slice is a logical network partition, defined within an operator network, that
can be dynamically created to meet certain SLA criteria (latency, reliability, throughput, geography).

Such virtual networks flow through the core network up to the optical network terminal or customer premises equipment. In other words, a full end-to-end virtual network is possible.

That means full control and autonomy is provided for each slice, with possibly-differentiated performance metrics for the network and services for the user of each slice. Network slicing is a product of the use of software defined networks and network functions virtualization (SDN/NFV)

In principle, network slices can be used by the network operator or any wholesale customer. The new solution is built around Nokia's cloud-native software platform Altiplano and open standards.  

Network slicing accomplishes in software much of what has traditionally been known as “wholesale,” where retail customers purchase the use of capacity from network owners.

Historically, such wholesale services have represented as much as 11 percent of total telecom service provider revenue, according to Ovum. That suggests revenue of perhaps $213 billion in 2021, Ovum predicts.

Essentially, Nokia says, network slicing creates a full Network as a Service (NaaS) offer for third parties that arguably provides a richer “own your own network” capability, compared to traditional spatial, spectral or temporal sharing techniques.

Business models might reflect historic preferences in each market. Some service providers, operating in markets where there is high reliance on wholesale, might extend NaaS using network slicing.

In other markets, where wholesale is less preferred, network operators likely will try and create new retail services that take advantage of network slicing.

source: nokia  

Nokia's programmable slicing solution creates virtual slices that look, feel and operate just like a physical network, the company says.

“Each service provider runs its own dedicated controller with a dedicated view of their slice of the network,” says Nokia.

That might well create new opportunities for any entities that formerly might have considered becoming a mobile virtual network operator or a private network operator in the fixed network realm.

Fiber to the Lightpole in 5G Era

“Fiber to the light pole” is one of the ways to think about optical fiber and other backhaul networks for 5G small cells.

Additional spectrum and smaller cell sizes are the two fundamental tools network designers can use to increase network bandwidth. In the pre-5G eras, when networks operated at lower frequencies, a macrocell tower (at 950 MHz) might transmit more than 17 miles, on flat terrain without major obstructions.

A 4G network using 1.8-GHz to 2.1-GHz signals might transmit only about 7.5 miles, by way of comparison. Low-frequency spectrum often is described as assets at and below 800 MHz (450 MHz, 600-MHz, 700 Mhz and 800 MHz). “Mid-frequency” tends to include 1.8 GHz to 2.1 GHz spectrum. “High frequency” traditionally has meant the 2.5-GHz range.

All that will change in the 5G era, as millimeter wave assets are commercialized. When millimeter spectrum is used (28 GHz, 39 GHz), small cells might cover a few to several hundred meters radius. In those cases, small cells might be placed about every other light pole along roads.

source: Cowen and Company

source: Cowen and Company

The following table shows the dependency of the coverage area of one cell on the frequency of a 3G network.

Frequency (MHz)	Cell radius (km)	Cell area (km2)	Relative Cell Count
450	48.9	7521	1
950	26.9	2269	3.3
1800	14.0	618	12.2
2100	12.0	449	16.2