Server Market, Metro and WAN Bandwidth Now Driven by Hyperscale Data Centers

source: Cisco

The global server market now is driven by huge data centers, much as global wide area network bandwidth demand is shaped by those same hyperscale data center sites, while metro traffic likewise is shaped by internal data center traffic and data center-to-data center connections.

"The real driver of global growth continues to be the hyper-scale datacenter segment," Gartner analyst Jeffrey Hewitt noted in a statement releasing the quarterly estimate. "The enterprise and small or mid-size business segments remain relatively flat as end users in these segments accommodated their increased application requirements through virtualization and considered cloud alternatives."

source: Cisco

OTT bypass--when OTT app providers terminate calls through an OTT app that have begun life as a normal fixed or mobile telephone call--could reduce carrier revenues as much as 30 percent, argues Andy Gent, Revector CEO.

OTT bypass affects carrier revenue because the bypass shifts call termination revenue away from telcos and to app providers.

“In the past month we have detected OTT Bypass across the globe with several operators reporting a reduction in termination call revenues of 25 percent, month on month,” said Gent.

“We have seen calls that have been made from one mobile phone number directly to another be received on a Viber app across three different continents and more than 15 network operators and national telecommunications regulators have contacted us regarding this issue in the past two weeks,”Gent said.

In the longer term, IDC expects server market growth to be driven by software-defined, disaggregated systems and network edge-deployed Internet of Things (IoT) compute," said Kuba Stolarski, research director, Computing Platforms at IDC.

By 2019, more than four-fifths (86 percent) of workloads will be processed by cloud data centers; 14 percent will be processed by traditional data centers. Much of that traffic will remain in single in metro areas, rather than crossing the wide area network.

Traffic between data centers is growing faster than either traffic to end-users or traffic within the data center, and by 2018,  traffic between data centers will account for almost nine percent of total data center traffic, up from nearly seven percent at the end of 2013, according to Cisco.

The portion of traffic residing within the data center will decline slightly between 2014 and 2019, accounting for 75.5 percent of data center traffic in 2014 and about 73 percent by 2019, Cisco predicts.

Global Data Center Traffic, 2014–2019
	2014	2015	2016	2017	2018	2019	CAGR 2014–2019
By Type (EB per Year)
Data center to user	613	760	946	1,185	1,495	1,886	25%
Data center to data center	234	321	432	564	723	905	31%
Within data center	2,602	3,342	4,233	5,235	6,358	7,566	24%
By Segment (EB per Year)
Consumer	2,103	2,758	3,550	4,444	5,599	6,885	27%
Business	1,346	1,665	2,061	2,540	2,977	3,472	21%
By Type (EB per Year)
Cloud data center	2,110	2,956	4,017	5,328	6,854	8,622	33%
Traditional data center	1,339	1,467	1,594	1,656	1,722	1,735	5%
Total (EB per Year)
Total data center traffic	3,449	4,423	5,611	6,984	8,576	10,357	25%

source: Cisco

source: Cisco

OTT Bypass a New Revenue Challenge for Mobile Operators

OTT bypass--when OTT app providers terminate calls through an OTT app that have begun life as a normal fixed or mobile telephone call--could reduce carrier revenues as much as 30 percent, argues Andy Gent, Revector CEO.

OTT bypass affects carrier revenue because the bypass shifts call termination revenue away from telcos and to app providers.

“In the past month we have detected OTT Bypass across the globe with several operators reporting a reduction in termination call revenues of 25 percent, month on month,” said Gent.

“We have seen calls that have been made from one mobile phone number directly to another be received on a Viber app across three different continents and more than 15 network operators and national telecommunications regulators have contacted us regarding this issue in the past two weeks,”Gent said.

Smartphone Apps Drive U.S. Content Usage

It will not surprise you that smartphones now are the ways most people interact with the Internet, in developed or developing nations and markets. In the U.S. market, for example, smartphones represent a plurality of Internet content use for nearly every age demographic.

comScore

Big Value of Industrial IoT is Protecting Capital Stock from Damage

source: Business Insider

Lower energy consumption is a value often cited for industrial Internet of Things applications, but protecting expensive equipment from potential damage arguably is the bigger driver of value in the industrial Internet of Things space.

Real time process control is the broad reason why edge computing (fog computing) is important, said David King, Foghorn Software CEO.

The reason process control is so important--beyond efficiency gains--is the value of protecting expensive equipment from damage and downtime.

In some cases, as with high-pressure water pumps, engineers might have only milliseconds to detect and prevent pressure conditions that cause cavitation and bubbles that damage a pump, said Sastry Malladi, Foghorn Systems CTO.

That is hard to do in a traditional cloud computing--rather than edge computing--environment. Communications latency alone can prevent proactive adjustments.

In manufacturing settings, $110,000 machines operate with specific winding tension, said Malladi. Equipment operators have only seconds to take machines offline when the winding tension is off, he added.

The point is that process control managers might have seconds to milliseconds to respond to a problem and modify operations to prevent equipment damage.

In other words, the value proposition includes better energy consumption, or more-efficient operations. But the real value is reducing damage to expensive equipment.

For Unconnected, "Killer App" Already Exists

source: Globalwebindex

Occasionally, one hears observers say that a “killer app” is required to get billions of people interested in using the Internet, assuming other barriers also are diminished (service cost, device cost, language, knowing how to use the Internet).

Others might say it is very clear that social networking is the killer app already in place. “Demand” is not the issue.

In other words, "end user demand" on the part of people who do not routinely use the Internet is not really lacking. People know what they want to do. We just have to build networks to deliver it, at prices people are willing and able to pay.

Network Interconnection Now is a Business Model Issue in India

Network interconnection is more than a technical process whereby companies connect their networks. The business rules for exchanging traffic can help some providers and harm other providers.

As Reliance Jio prepares to enter the India mobile market, interconnection rules once again are highlighted as a material contributor to firm revenues and business models. The Telecom Regulatory Authority of India, for example, is taking a look at termination charges.

There is some possibility TRAI could recommend essentially eliminating interconnection charges. That would reduce revenue for the largest mobile companies in India, and reduce cost for Reliance Jio.

The reason is simply that when networks of unequal size exchange traffic, the large networks always will, on balance, terminate many more calls than the small carriers.

At the same time, the small networks will mostly originate traffic, and will get relatively small amounts of terminating traffic from the big networks.

So low or zero-rated termination helps small carriers, and hurts big carriers, since it is the big carriers that terminate most traffic.

If Reliance Jio plans to offer “no incremental cost” (“free”) domestic calling, then interconnection charges to terminate those calls on other networks has to be set at “no incremental cost” levels as well, at a high level, or else Reliance Jio will subsidize every minute of use.

Reliance Jio Will "Crash Prices" and Drive Some Competitors from the Market

Two developments are virtually certain as Reliance Jio enters the India mobile market: revenue will “crash” and markets will consolidate.

Case studies of seven markets over the last 10 years revealed that whenever a "radically new disrupter" came in it has almost always led to two things: first, crash in voice service prices as consumers make a shift to higher data usage and second, demise of weaker players, an analyst said.

Beyond that, the overall trend in the global telecom business for three decades has been lower prices, greater competition and differentiated business strategies. Over time, the pricing trend for telecommunications products has had a tendency to move towards zero, or marginal cost pricing.

The clear problem with marginal cost pricing is that the supplier only recovers the incremental cost of producing the extra units, not the sunk cost of the infrastructure.

In principle, marginal cost pricing assumes that a seller recoups the cost of selling the incremental units in the short term and recovers sunk cost eventually.

The growing question is how to eventually recover all the capital invested in next generation networks, if retail pricing moves to “marginal cost.”

Indeed, some already argue that tier one telcos do not recover their cost of capital, perhaps an indication that marginal cost pricing is dangerous to the long term health of the industry. .

As a rule, any industry touched by Internet distribution tends to see a trimming of supplier profit margins. In fact, that is an important strategy for digital disruptors, where the strategy literally is to destroy profit margins in a traditional business, gaining share and then dominating the new business, with permanently lower profit margins, and possible lower gross revenues.

That is the theory that underpins the pursuit of “zero billion dollar markets.” One sense of the word is that big markets get created when whole new industries are founded. But one other use is more ominous for incumbents.

That is reliance on marginal cost pricing to literally “destroy” the pricing regime in an existing market, allowing a new competitor with radically lower cost structure to displace the current leaders. That is the essence of the phrase “analog dollars, digital dimes and mobile pennies.”

More Regulation for Skype, WhatsApp in EU?

source: Ali Saghaeian

Even if one believes that telecommunications policy should treat all similar services provided by similar entities in a similar way, there are two basic approaches to achieve those ends.

Regulators can lessen rules on incumbents, to match rules applied to new competitors, or apply incumbent rules to new providers. In the European Union, it appears telecom regulators are preparing to apply some rules to over-the-top voice and messaging apps.

source: Ali Saghaeian

Incumbent telecom providers have argued for nearly two decades that over-the-top voice and messaging services such as Microsoft’s Skype and Facebook’s WhatsApp are functional substitutes for carrier voice and messaging, and should be covered by the same rules applied to carrier voice.

It can be argued that OTT messaging and voice are imperfect, or only partial substitutes, with limitations. Skype is further along the process of supporting communications on an “any-to-any” basis. WhatsApp still remains a social app, requiring all users to join the community to use the app.

On the other hand, the range of functionality keeps increasing.

Skype may have to offer emergency-calling services for European customers. In cases where it assigns users telephone numbers—allowing for the receipt of calls from traditional phones—it could also be required to let those users take their numbers with them if they decide to move to a different provider.

WhatsApp could be subject to new rules on network security.

Whether one agrees that OTT apps are full functional substitutes or not, regulators have a choice: lighten burdens or increase them, to “level the playing field.”

Some of us would argue that less regulation makes sense for products in decline, as characterizes the voice and messaging services business.

5G Will Bring Differentiated "Mobile" and "Not Mobile" Tariffs

If fixed wireless and 5G begin to compete for Internet access accounts now served by fixed networks, several key business model changes will be required. First, speeds will have to increase by an order of magnitude or perhaps two orders of magnitude.

Prices also will have to drop, on a cost-per-megabyte basis, as mobile bandwidth has historically been priced significantly higher than fixed network bandwidth.

Another likely development for mobile rate plans is differentiated pricing for non-mobile access, relying either on Wi-Fi offload or small cell pricing when users are not mobile, in large part because the huge amounts of new bandwidth to support 5G will be available mostly in small cell areas.

Just how big a difference can be gleaned by the amount of capacity available below 1 GHz and newer spectrum in the gigaHertz ranges. Basically, all presently-deployed mobile communications spectrum represents less than one gigaHertz of total spectrum. New allocations for 5G will add scores of gigaHertz of new spectrum.

Bandwidth requirements will be shocking. As much as 56 GHz of 5G spectrum might be needed, one study of the United Kingdom by Real Wireless suggests. Even if that is wildly off base, otehr researchers still believe that, to support 5G, 3.1 GHz per mobile operator might be required, according to Samsung researchers, with consumption as high as 12.75 GHz per user in small cells.

By way of comparison, the typical tier-one mobile operator in European markets uses about 590 MHz of spectrum, each.

All that suggests differentiated pricing for "mobile" bandwidth, compared to non-mobile use.

source: Telecoms.com

source: policytracker

Small Cell Bandwidth Per User Could Reach 12.75 GHz for Each User

Though mobile spectrum for purposes of coverage might not change nearly as much, small cell bandwidth could mean consumption as high as 12.75 GHz per user, in the small cell areas.

That could mean, by some estimates, a need for as 56 GHz of spectrum, to support multiple service suppliers, one study of the United Kingdom by Real Wireless suggests.

For 5G, 3.1 GHz per mobile operator might be required, according to Samsung researchers.

The typical tier-one mobile operator in European markets uses about 590 MHz of spectrum, each.

source: Samsung

Are FTTH Take Rates Really 50%?

According to the Fiber to the Home Council, half of homes passed by fiber actually buy service.

That seems too high, and is almost certainly a statistical artifact caused by a Verizon sale of assets that subtracted millions of copper-served homes from the actual “homes passed” base.

Verizon supplies most FTTH connections in the United States, so any big change at Verizon will affect the whole market (AT&T’s FTTH passings are growing, and other ISPs operate, but Verizon is the dominant provider of fiber-to-home connections.

In the first quarter of 2016, Verizon FiOS Internet access connections reached 7.1 million accounts, on a base of about 15 million homes. So adoption of FiOS Internet services could be as high as 47 percent.

It all depends on how many homes passed Verizon has. Prior to asset sales to Frontier Communications, there were 19.8 million homes passed by FiOS networks, so take rates for customers able to buy it were once about 36 percent.

Then Verizon sold 4.8 million lines (and more homes passed than that) to Frontier Communications. After the asset sales, Verizon now passes about 15 million homes.

In other words, because Verizon sold assets that mostly did not have FTTH activated, the denominator was reduced more than the numerator when calculating fiber adoption. But nothing really changed in terms of Verizon adoption rates or availability.

It is correct to say that Verizon FTTH take rates (Internet and video) are about 50 percent, where the services are available for purchase.

Still, it has to be noted that other competitors will find it hard to match those levels of adoption. Verizon FiOS has been marketed the longest in the U.S. market, and generally has faced access competition primarily from cable operators. If Verizon gets 47 percent adoption, then cable could potentially get 53 percent.

We will see what happens as competition grows, especially as Comcast activates gigabit capabilities that operate faster than FiOS. Eventually, we also should see additional fixed wireless and mobile competition, plus potential independent ISP market entry in a few instances.

New Zealand Launching Gigabit Services Nationwide

Chorus, the New Zealand wholesale access provider,  is extending its gigabit (1Gbps) residential and business fiber Internet access service across its entire “Ultra-Fast Broadband” (UFB) footprint, starting October 1, 2016.

Currently, the average download speed across Chorus' networks is 30.5 Mbps.

In practice, customers will see download speeds between 900 Mbps and 970 Mbps and upload speeds of up to 500Mbps.

Chorus’ residential wholesale gigabit broadband service will be available to broadband retailers at an introductory price of $60 per month until 30 June 2017, after which it increases to $65 per month.

The business service will be priced at $75 per month from launch.

4K is Marketing Hype for TVs and Smartphones, Useful Mostly for Tablet, PC Applications

Up to this point, most advances in image quality (high definition TV, for example) have been about “large screen” viewing. Ironically,  4K and higher resolution video standards (though undoubtedly propelled by large and small screen sales) might actually be the first image improvement standard with greater relevance for tablet or PC screens or “lean forward” applications than traditionally has been the case.

source: rtings.com

Whether there are perceivable advantages for smartphones and TVs is questionable.

The reason is that most consumers, sitting at normal TV viewing distances in their living rooms, media rooms or family rooms, will not actually be able to perceive a quality improvement over HDTV.

Image improvement really requires that users be very close to their screens. In fact, for HDTV, viewers have to avoid sitting too close.

In fact, 4K Ultra HD resolution is not worth it if you are sitting more than six feet away and have a 50-inch TV, since the human eye cannot tell the difference between 4K resolution and HDTV.

One additional caveat: screen sizes are quoted on the diagonal. TV viewing distances might, or might not, refer to actual vertical height.

So “ultra HD” only makes sense if you want a really big screen and plan on sitting close to it. Or, perhaps, if your vision vision actually is “perfect.”

source: appliancesonline

At around the distance of 10 inches, people with average vision can discern an image with ~344 PPI, and people with perfect vision can discern an image with higher resolution up to ~573 PPI.

If one assumes that most people cannot detect an image quality difference greater than 344 PPI, 4K is wasted on any screen size smaller than about eight inches.

Screen Size	Optimal Distance 1080p
25"	3.3' (1 m)
30"	4' (1.22 m)
35"	4.6' (1.40 m)
40"	5.3' (1.62 m)
45"	6' (1.83 m)
50"	6.6' (2.01 m)
55"	7.3' (2.23 m)
60"	8' (2.44 m)
65"	8.6' (2.62 m)
70"	9.3' (2.83 m)

EU Ponders More Regulation for OTT Voice, Messaging

How to regulate new services and apps, how to foster innovation and yet maintain a “level playing field” always are key issues in the telecommunications business, exemplified most clearly by the debates policymakers, app and service providers have had for decades over IP-based voice services.

In India, mobile operators have argued that over the top voice and messaging apps are, in fact, full substitutes for traditional mobile voice and messaging, and therefore should be covered by the same rules as carrier voice and messaging.

The European Union, for its part, is considering mandating some security rules  for WhatsApp, Skype and Apple Inc's FaceTime.

The proposed rules would increase levels of regulation for OTT apps.

Sometimes, aside from the sheer merits of regulating competitors, there are industrial policy issues as well.

“The proposal is part of a broader drive to level the playing field between European companies and mainly U.S. tech firms,” Reuters reports.