11% More Hyperscale Data Centers Added in 2018

The number of large data centers operated by hyperscale providers rose by 11 percent in 2018 to reach 430 by year end, says Synergy Research Group. And that could continue to reshape wide area network communications, as most global traffic now is generated by the hyperscale data centers.

Amazon, Microsoft, Google and IBM each have 55 or more data center locations with at least three in each of the four regions of North America, APAC, EMEA and Latin America.

Most  data traffic these days is generated by cloud computing, but most of the actual data communications--as much as 72 percent--happens within data centers, according to Cisco.

source: Cisco

As you likely would guess, hyperscale data centers represent a large portion of overall data, traffic, and processing power in data centers, accounting for 34 percent of total traffic within all data centers and driving 53 percent of in-data-center traffic by 2020.

Hyperscale data centers will also represent 57 percent of all data stored in data centers and 68 percent of total data center processing power.

At the same time, the huge amount of wide area network traffic between hyperscale data centers is causing disintermediation (substitution, or taking out the middleman) in the wide area networking business.

More than half of all WAN traffic now moves on private networks built and operated directly by enterprises, especially app, commerce and content providers.

source: TeleGeography

Alibaba and Oracle also have a notably broad data center presence. The remaining firms tend to have their data centers focused primarily in either the United States (Apple, Facebook, Twitter, eBay, Yahoo) or China (Baidu, Tencent).

In 2018 the Asia-Pacific and Europe-Middle East-Africa  regions featured most prominently in terms of new data centers that were opened.

Among the hyperscale operators, Amazon and Google opened the most new data centers in 2018, together accounting for over half of the total.

source: Synergy Research Group

Bell Canada Wants to Mine its Customer Behavior Data

North American telcos (Canadian and U.S.) historically have faced more-stringent privacy rules and limitations on data mining than have applied to app providers such as Google, Facebook and other online app providers.

So it perhaps comes as no surprise that Bell Canada began asking its customers in December for permission to track everything they do with their home and mobile phones, internet, television, apps or any other services they get through Bell or its affiliates.

And that is one of the key issues for any telco that wants to mine its customer data for insights that can be used to build a targeted advertising business.

Key to the effort, apparently, is avoidance of divulging any personally-identifiable information. "Bell's marketing partners will not receive the personal information of program participants; we just deliver the offers relevant to the program participants on their behalf," Bell Canada says.

The big upside, some believe, is the ability to earn revenue by making targeted advertising available to the Bell Canada customer base, not just properties or content that Bell Canada might own.

Xandr and Verizon Verizon Media Group provide more evidence that at least some telcos believe they can leverage data to support advertising revenue streams. Just how successful they might eventually be is yet open to question.

In principle, mobile service providers and telcos have multiple data stores they could work with, provided they have customer permission to do so: video behavior, internet access behavior, mobile phone and perhaps fixed network calling behavior.

In the U.S. market, for example, it appears internet service providers can record and sell customer browsing history, data on which apps and services are used. How valuable that might be is among the questions one might ask, in the same way that one might question the value of call detail, in an era where most of the insight is related to internet content behavior.

Telcos process and possess lots of data. But some might question how much data telcos actually can use to support advertising and marketing services, given privacy regulations and the sorts of user actions they can track.

There seems to be less doubt about ability to use some of that data to improve operations. AT&T has mined data for years to wring business value on the operations side of the business (network management, fraud detection, perhaps churn management, customer service). What seems less clear is whether there is significant revenue-affecting upside.

Obviously call detail records, mobile location and data related to customer use of internet access services is the potential mine for revenue-generating or revenue-impacting value.

The issue is how much granularity is available for the internet and app use parts of the data mine, compared to calling or text message behavior; how long such data can be retained and analysed; and how the insights can be applied to advertising or other revenue-generating activities.

Why Can Mobile Leverage Its Data to Add Revenue?

People tend to understand that mobile and other service providers have mined their data to improve operations. The next question is where, and whether, they can leverage customer data to create revenue upside. 

Xandr and Verizon Verizon Media Group provide evidence that at least some telcos believe they can leverage data to support advertising revenue streams. Just how successful they might eventually be is yet open to question, as app providers lead the digital advertising business.

In the U.S. market, for example, it appears internet service providers ISPs can record and sell customer  browsing history, data on which apps and services are used. How valuable that might be is among the questions one might ask, in the same way that one might question the value of call detail, in an era where most of the insight is related to internet content behavior.

Telcos process and possess lots of data. But some might question how much data telcos actually can use to support advertising and marketing services, given privacy regulations and the sorts of user actions they can track.

There seems to be less doubt about ability to use some of that data to improve operations. AT&T has mined data for years to wring business value on the operations side of the business (network management, fraud detection, perhaps churn management, customer service). What seems less clear is whether there is significant revenue-affecting upside.

Obviously call detail records, mobile location and data related to customer use of internet access services is the potential mine for revenue-generating or revenue-impacting value.

The issue is how much granularity is available for the internet and app use parts of the data mine, compared to calling or text message behavior.

In developed countries, most people are online daily. By some estimates, 78 percent of customers are online everyday, but telcos typically reach less than one percent of people digitally on a daily basis.

Most Europeans use the mobile internet daily. But one might question how much of the total digital footprint is accessible to mobile operators, compared to app providers, mobile phone suppliers, operating system and other participants in the internet ecosystem.

It would be fair to say that Xandr and Verizon Media Group have some obvious advantages, including direct customer relationships and advertising inventory (linear video and digital content assets).    

To oversimplify ownership of digital footprint, Google knows your location, the content of your emails, your search terms, in some cases your contact list and calendar, and websites you visit.

Facebook knows who your friends are and has some idea of what you like. Amazon knows what you buy. So do credit card companies and other financial institutions.

So as connectivity providers look for some way to leverage their own digital footprint assets, what have they got? Your location; who you call and text, who calls you, and sends you messages; what websites you visit and some high-level details about what content you consume.

In addition, AT&T, the largest provider of linear video subscriptions in the U.S. market, has local advertising inventory that it can sell, as well.

How valuable that data might be, and how it can be used, are questions, though. Up to this point, many telco data mining efforts have focused on operating problems such as fraud, not revenue generation.

Generally there are limits on mobile data retention that limit how long such data can be kept, as well as laws on use of that data. But privacy rules might prevent the mining and application of such data.

That noted, some telcos believe they can leverage data to support advertising businesses, especially related to local tV advertising, using set-top-box viewing data, for example.

Perhaps the more challenging issues are ways to monetize mobile traffic, web browsing and app usage, supplying insights to potential partners, for example. In a smaller number of cases, mobile operators might own content assets where insights can be applied directly.

Greater restrictions are likely going to be considered on use and retention of data possessed by application providers as well.

The point is that mobile service providers--in principle--might be able to leverage location, calling circle and website visit details to create algorithms useful when building an advertising business, for example, consistent with privacy rules that seem to be getting more stringent.

But there are some important limitations. Mobile operators capture three main kinds of information: devices connected to the network, metadata about the packets of data that run through the network, and information about the content contained in the packets being downloaded or uploaded.

But most of that data is not personally identifiable. That might limit the ability to target advertising or content to specific customers, but does allow for creation of larger algorithms about behavior in general.

One might question the value of metadata (details such as the origin and destination of the packet, whether or not the packet contains data from a real-time service like VoIP, and the amount of data in the packet.

The header gives the operator a rough idea of what the content is for, without disclosing any actual details of the content itself.

Still, it is hard to envision, at the moment, how much of the mobile and digital footprint Xandr and Verizon Media Services might be able to leverage, aside from offering local advertising inventory in a relatively traditional sense.

"New and Improved" Works for 4K, Gigabit Internet Access and Most Other Products

Most of us are used to the idea that computing and other consumer devices tend to get better every year or two. And while some devices, such as smartphones, might also become more expensive as they become more powerful, other devices, such as TVs, tend to get cheaper as they add functionality.

To anticipate an objection, most of us would not object to having a bigger TV screen, with better resolution, even if in fact we do not use such devices in ways that always allow us to take advantage of higher resolution, and so long as there is no price disadvantage when buying such a display.

That is the same sort of issue we often encounter when buying an internet access service. Many consumers buy services that actually do not provide any visible advantage in terms of user experience.

That often holds even for their TVs, which increasingly sport higher-definition modes (4K, for example).

So here is the issue: A TV screen might be capable of displaying 4K content, but that might not help if a consumer’s video inputs or program suppliers do not support 4K resolution. That can be the case for an over-the-air broadcast, a linear video feed or streaming video input.

There is another common scenario as well, where the TV and the input can support 4K, and some 4K content is available, but the viewer is sitting far away that the human eye cannot perceive the higher resolution.

Looking at a 50-inch screen 8K TV screen, the viewer has to sit 1.7 feet from the screen to benefit from the resolution. Keep in mind: that is about the distance most people sit from a PC screen. Few people--if any--are ever going to sit that close to a TV screen.

Many people sit about nine feet away from their TV. You’d need between a 275” and 280” 8K screen to push the optimal viewing distance out to nine feet, some would note.

The optimal distances for a 65” screen are two feet for 8K and 4.3 feet for 4K. How many people will sit that close to a TV screen? Virtually nobody.

If you sit somewhere between these two distances, you will see some, but not all, of the added detail in the 8K image. If you sit 4.3 feet or further from the screen, the 8K and 4K images will look identical and you wasted your money on 8K, experts say.

The human eye, like any lens, cannot infinitely resolve the difference between two lines or two points at any distance. The Reyleigh formula mathematically describes lens resolution. In other words, if you are looking at something very close, you can see many details that are not visible at greater distances.

When it comes to televisions touting new 4K technology, "a regular human isn't going to see a difference," said Raymond Soneira, head of display-testing firm DisplayMate Technologies.

And distance matters for humans who watch TV, as do screen sizes. Basically, as resolution increases, viewers have to sit closer to screens to see the better resolution images.

And that is where furniture makes a difference. Most people do not actually sit close enough to their TV screens to benefit from higher resolution. And the higher resolution gets--replacing a 1080p screen with 4K, then 8K--the closer one has to sit.

source: Carlton Bale

The optimal viewing distance is the distance at which the eye can resolve all the detail in the image without being able to see the pixels on the screen. Assuming a viewer has 20/20 vision, and is looking at a 50-inch TV screen, the person has to sit 3.2 feet from the screen to benefit from 4K resolution.

Screen size	Optimal 8K distance	Optimal 4K distance	Optimal 1080p distance
75"	2.5 ft. (0.79 m)	4.9 ft. (1.48 m)	9.7 ft. (2.97 m)
70"	2.3 ft. (0.70 m)	4.6 ft. (1.40 m)	9.2 ft. (2.79 m)
65"	2.0 ft. (0.61 m)	4.3 ft. (1.31 m)	8.6 ft. (2.62 m)
60"	2.0 ft. (0.61 m)	4.0 ft. (1.22 m)	7.7 ft. (2.36 m)
55"	1.7 ft. (0.52 m)	3.7 ft. (1.14 m)	7.2 ft. (2.18 m)
50"	1.7 ft. (0.52 m)	3.2 ft. (0.96 m)	6.6 ft. (2.01 m)
43"	1.4 ft. (0.44 m)	2.9 ft. (0.87 m)	5.7 ft. (1.75 m)
40"	1.4 ft. (0.44 m)	2.6 ft. (0.79 m)	5.2 ft. (1.57 m)
32"	1.1 ft. (0.35 m)	2.0 ft. (0.61 m)	4.3 ft. (1.31 m)

source: Forbes

Even if the necessary technology was in place to watch and stream 8K, 8K content was plentiful, and prices dropped to reasonable levels, buying an 8K TV would make no sense whatsoever for most people who watch TV in their homes, others argue.

Why not? The human eye cannot resolve the level of detail that’s present in an 8K image at the distance most people sit, or would want to sit, from their TV.

PC monitors and small screens such as smartphones and tablets, on the other hand, are venues where much-better resolution does help. Resolution of 4K does provide benefits on monitors of 20 inches, for people with 20/20 vision.

For the average healthy adult with 20/15 vision, you should be able to easily tell the difference between a 2K and 4K 15.6" screen at a distance of 22 inches.

If the choice is between a 1080p screen and a 4K screen on a notebook PC, 4K will be noticeably better at a viewing distance of 32.5 inches or closer.

Is "GHz" a Measure of Capacity, Frequency or Clock Speed? Yes.

Bandwidth (capacity) discussions sometimes get confused because we use some terms in different ways. Consider the ways we use “Hertz.” We talk about signal frequencies as a way of describing different blocks of capacity at different wavelengths (optical or radio frequency).

source: Simplifying Theory

Think of this usage as musical notes, which can range from low to high in multiple octaves. So mobile signals can use spectrum at lower (600 MHz, 800 MHz), mid-range (2 GHz) or high frequencies (30 GHz to 100 GHz)

We also use the same terms to describe channel sizes, which relate to potential data rates.

source: Bitfeed

Personal computer makers use similar terms to describe processing speed (clock speed).

Then we use related terms such as “megabits per second” (Mbps) to describe throughput and “megabytes” as a measure of capacity or throughput as well.

Here’s a tutorial on such things, at least in the areas of signal frequencies and channel bandwidth.

Satellite Connectivity Providers Challenged by Shift Away from Linear Video

The satellite services industry faces challenges similar to those faced by other connectivity providers, namely a shift of demand. In 2017, 75 percent of service provider revenues were generated by delivering video entertainment (television).

To the extent that video distribution is moving towards an on-demand mode, that threatens the primary revenue source. In that respect, the satellite services industry faces the same problem as do providers of linear subscription television.

Geosynchronous satellites, though, have a major limitation where it comes to on-demand services, namely latency. New constellations of low earth orbit satellites will address that issue, but one can understand the industry concern about future revenues.

When voice (first long distance revenues, then calling revenue generally) revenue declined, first landline, then mobile operators,  were forced to find big new revenue substitutes. So no less than mobile and other service providers, satellite connectivity suppliers will seek new roles in internet of things and other on-demand services.

source: Bryce