Hard to Separate Edge Computing, 5G, Optical Networking, AI and VR

In terms of evaluating potential impact, it is becoming very difficult to separate 5G and optical networking from edge computing from applied artificial intelligence as platforms to create new use cases and applications driving value for end users and enterprises as well as revenue opportunities for computing and communications providers. 

Consider the new industry forum, the Innovative Optical and Wireless Network Global Forum, founded by Nippon Telegraph and Telephone Corporation, Intel Corporation and Sony Corporation.

The global forum’s objective is to accelerate the adoption of a new communication infrastructure that will bring together an all photonics network infrastructure including silicon photonics, edge computing and connected computing.

That includes artificial intelligence, dynamic and distributed computing, as well as digital twin computing (a computing paradigm that enables humans and things in the real world to be re-created and interact without restrictions in cyberspace).

For a few decades now, digital simulations and environments  have been touted as new ways to use “virtual worlds” to support real world commerce. Some might recall SimCity, a 1989 videogame. So virtual reality is not essentially new. Digital twinning arguably is new. 

The concept of creating virtual representations of real worlds now is referred to by some as digital twin computing, and applies mostly in industrial settings. A digital twin is a dynamic, virtual representation of a physical asset, product, process, or system, human or non-living entity.

Such digital representations model the properties, condition, and attributes of the real-world counterpart, and originally were useful for simulating the capabilities of machine tools in a safe and cost-effective way, as well as identifying the root causes of problems occurring in physical tools or infrastructure. 

If a physical machine tool breaks down or malfunctions, engineers can evaluate the digital traces of the digital twins’ virtual machines for diagnosis and prognosis.

Some believe the use of digital twinning, creating virtual replicas of physical objects and devices, eventually including humans, will be an important outcome of network and computing advances. 

Twinning, AI in general and virtual reality, in turn, will often going to be related to edge computing. “Forty-three percent of AI tasks will be handled by edge computing in 2023,” said Kwon Myung-sook, CEO of Intel Korea, during a forum in Seoul. “AI devices empowered with edge function will jump 15-fold.”

Azure now allows users to invoke supercomputer instances from their desktops. Nvidia markets an edge computing platform incorporating AI, for example. 

Google Fiber Now Sells Gigabit Internet Access as its Sole Offer

When Google Fiber launched its symmetrical gigabit internet access service, typical U.S. fixed network internet access speeds might have averaged about seven to nine megabits per second. In retrospect, that was about the time that U.S. access speeds began a nonlinear ascent, powered by U.S. cable operator speed boosts. 

It is worth noting that fiber-to-home average speeds were not much better, about 7.7 Mbps downstream, with cable hybrid fiber coax average speeds in the 5.5 Mbps range and digital subscriber line lagging with an average speed of about 2.2 Mbps downstream. 

So Google Fiber, offering a symmetrical gigabit connection for $70 a month, was quite disruptive, in terms of reshaping expectations about speed and price per Mbps. 

Of course, perhaps we should not have been surprised at the growth of internet speeds. Back when dial-up modems were running at 56 kbps, Reed Hastings, Netflix CEO,  extrapolated from Moore's Law to understand where bandwidth would be in the future, not where it was “right now.”

“We took out our spreadsheets and we figured we’d get 14 megabits per second to the home by 2012, which turns out is about what we will get,” says Reed Hastings, Netflix CEO. “If you drag it out to 2021, we will all have a gigabit to the home." 

So far, internet access speeds have increased at just about those rates. So what is the meaning of Google Fiber dropping its lower-speed tiers and selling only one service: a symmetrical gigabit service for $70 a month?

There are several likely reasons for the switch. The most obvious is that Google Fiber now routinely competes against cable TV operators offering speeds of a gigabit at prices that are not so different from Google Fiber, in real terms, once bundles and other promotions are included. 

Just as important, Google Fiber’s 100-Mbps service, sold for $50 a month, might actually be more expensive than some cable offers when internet access is purchased as part of a bundle. So one reason for streamlining is simply that the 100-Mbps offer is not getting traction, because it no longer offers value, compared to standard cable TV internet access offers, though it still often does when comparing telco services using copper access plant. 

But Google Fiber does not, as a practical matter, compete against slow DSL, but against cable TV services, as cable has about 66 percent of the installed base of customers, on average, across the United States, and has been from telcos continuously since perhaps 1999. 

Over the last 20 years, it would be hard to find a single year where cable broadband net account gains were not about 60 percent to 70 percent of all net gains, and over the last decade virtually all the net gains. 

To be sure, only about four percent of U.S. fixed network internet access customers seem to buy a gigabit per second service from any internet service provider. Altogether, some 77 percent of U.S. households buy internet access running between 50 Mbps and 300 Mbps.

In fact, 100 Mbps might have become the average U.S. downstream speed in 2018. 

The point is that Google Fiber and other independent ISPs selling gigabit per second service offer the most-distinctive value proposition as providers of symmetrical gigabit services, priced around  $70 to $90 per month. Comcast generally prices stand-alone gigabit service between $105 and $140 a month. 

Comcast has said that 75 percent of its customers now buy services operating at 100 Mbps or faster. 

The point is that Google Fiber no longer is the sole gigabit offer in most U.S. urban or suburban markets and likely finds few customers interested in its 100-Mbps offer. Google Fiber’s uniqueness in most markets is its symmetrical gigabit offer, since cable services remain asymmetrical. 

By 2023, 1/2 of Hyperscale Data Center Capex Might be Going to the Edge

Alphabet, Alibaba, Amazon, Apple, Baidu, Facebook, Microsoft, Rakuten and Tencent, which operate the biggest hyperscale data centers, also now are making initial moves to “own the edge” as well. Though in 2019 they might collectively spend about five percent of total capex on edge computing, by 2023 they could be devoting as much as 50 percent of capex on edge computing, according to a forecast by Technology Business Research. 

source: Technology Business Research

The obvious question for connectivity providers is whether, and how much, new potential “partnerships” with some of the hyperscale providers might represent value and revenue for connectivity providers. 

Up to a point, edge computing facilities will scale faster if partnerships can be struck between ecosystem participants (central office owners, cell tower owners, connectivity providers and data centers, for example). 

The tricky task is striking deals that spread revenue in ways seen as fair by the participants. And, of course, eventual channel conflict seems somewhat inevitable. That noted, the multi-cloud trend and market for edge data center space and support (from racks to compute cycles) might create several niches. 

AWS already has signed Verizon, Vodafone, KDDI and SK Telecom as access partners, and also appears to be leasing data center space from the carriers. So, at the very least, it appears each of the telco partners will earn some data center leasing revenue from AWS, which is placing its servers in telco racks (presumably often in central offices). 

The telco partners likely also gain some amount of dedicated access (capacity) revenue from AWS and any enterprise customers using the AWS edge computing service. 

There is some potential indirect benefit, as often is the case, even from partnerships with little direct incremental revenue, as the availability of AWS edge computing services makes the connectivity service more valuable. 

The AWS partnership does provide near-term value in allowing each of the telco partners to tout their presence in the edge computing market, at low investment costs to them. 

On the other hand, it is hard to see much incremental revenue upside from this sort of partnership. As always, the biggest revenue and profit returns will flow to providers that own the infrastructure and the services, as always is the case for owner’s economics.

Still, when markets are young, revenue upside is likely small and investment costs high, that “creep in” at low cost strategy reduces risk. Telcos gain some marketing platform advantages and some “rent rack space” revenue.

Greater levels of investment and risk come with business models “up the stack,” including operating an owned “edge computing facilities as a service” business. Another step up is to sell edge computing capabilities directly, not simply selling colocation space to third parties.

More intense models including some forms of system integration, owned business-to-business apps, and finally, direct sale of full retail apps to end users. Those sorts of efforts have been difficult for telcos in the past, but are not impossible. 

On the other hand, it is hard to see how edge computing becomes a significant revenue generator for telcos unless they participate at some level beyond selling rack space.

Will Telcos Eventually Move Beyond AWS Wavelength?

AWS partnerships with a number of telcos as part of Wavelength, one of several new AWS edge computing initiatives,  does inevitably raise the issue of how “partnerships” represent value and revenue for connectivity providers. 

AWS already has signed Verizon, Vodafone, KDDI and SK Telecom as access partners, and also appears to be leasing data center space from the carriers. So, at the very least, it appears each of the telco partners will earn some data center leasing revenue from AWS, which is placing its servers in telco racks (presumably often in central offices). 

Each telco probably also gains some dedicated access (capacity) revenue as well.

There is some potential indirect benefit, as often is the case, even from partnerships with little direct incremental revenue, as the availability of AWS edge computing services makes the connectivity service more valuable. 

The AWS partnership does provide near-term value in allowing each of the telco partners to tout their presence in the edge computing market, at low investment costs to them. 

On the other hand, it is hard to see much incremental revenue upside from this sort of partnership. As always, the biggest revenue and profit returns will flow to providers that own the infrastructure and the services, as always is the case for owner’s economics.

Still, when markets are young, revenue upside is likely small and investment costs high, that “creep in” at low cost strategy reduces risk. Telcos gain some marketing platform advantages and some “rent rack space” revenue.

Greater levels of investment and risk come with business models “up the stack,” including operating an owned “edge computing facilities as a service” business. Another step up is to sell edge computing capabilities directly, not simply selling colocation space to third parties.

More intense models including owned business-to-business apps, and finally, direct sale of full retail apps to end users. 

Wavelength limits telco risk, but also telco revenue and strategic upside. Eventually, some may try--or expand--other initiatives with more revenue and profit upside. Longer term, at least some of the connectivity providers may attempt to enter or grow other edge-related businesses. Both AT&T and Verizon own businesses in the auto communications area, for example. 

Still, AWS is making its edge computing strategy clearer, launching several initiatives, showing the competition access providers will face from others in the ecosystem.

AWS Wavelength embeds AWS compute and storage services within telecommunications provider data centers at the edge of the 5G networks.

AWS Local Zone extends edge computing service by placing  AWS compute, storage, database, and other select services closer to large population, industry, and IT centers where no AWS Region exists today. 

AWS Local Zones are designed to run workloads that require single-digit millisecond latency, such as video rendering and graphics intensive, virtual desktop applications. Local Zones are intended for customers that do not want to operate their own on-premises or local data center.

Likewise, AWS Outposts puts AWS servers directly into an enterprise data center, creating yet another way AWS becomes a supplier of edge computing services. “AWS Outposts is designed for workloads that need to remain on-premises due to latency requirements, where customers want that workload to run seamlessly with the rest of their other workloads in AWS,” AWS says.  

AWS Outposts are fully managed and configurable compute and storage racks built with AWS-designed hardware that allow customers to run compute and storage on-premises, while seamlessly connecting to AWS’s broad array of services in the cloud.

AWS Launches Three Edge Computing Services

At least some tier-one connectivity providers will try to create edge computing as a service facilities and businesses. In that effort, they will possibly face competition from other potential competitors, including tower companies, the hyperscale computing giants, data centers and possibly new entrants as well.

AWS is making its edge computing strategy clearer, launching several initiatives. AWS Wavelength embeds AWS compute and storage services within telecommunications provider data centers at the edge of the 5G networks.

AWS Local Zone extends edge computing service by placing  AWS compute, storage, database, and other select services closer to large population, industry, and IT centers where no AWS Region exists today. 

AWS Local Zones are designed to run workloads that require single-digit millisecond latency, such as video rendering and graphics intensive, virtual desktop applications. Local Zones are intended for customers that do not want to operate their own on-premises or local data center.

Likewise, AWS Outposts puts AWS servers directly into an enterprise data center, creating yet another way AWS becomes a supplier of edge computing services. “AWS Outposts is designed for workloads that need to remain on-premises due to latency requirements, where customers want that workload to run seamlessly with the rest of their other workloads in AWS,” AWS says.  

AWS Outposts are fully managed and configurable compute and storage racks built with AWS-designed hardware that allow customers to run compute and storage on-premises, while seamlessly connecting to AWS’s broad array of services in the cloud.

Mobile Execs Say OTT Competition is Greatest Challenge

Mobile operator executives continue to believe that “disruptive competition” from over-the-top app providers is the single greatest business challenge they face, followed by uncertainty of the regulatory environment, lack of organizational agility and return on investment.  

Given the shrinking sales volume and revenue from legacy revenue sources (voice, messaging) challenged profit margins on newer services (internet access, enterprise data) and difficulty of competing in some new areas (data centers), such concerns are understandable. 

source: EY

Optimists continue to believe that connectivity revenues can benefit from “supporting” OTT app providers. And, of course, there is truth in that belief. The internet access revenue stream would not exist were it not for the value of internet-based apps. 

Connectivity providers that own content services, enterprise unified communications services or other OTT services also benefit. 

On the other hand, the internet also enables several types of competition for connectivity provider services, in the form of product substitutes. 

It remains unclear how the “edge computing as a service” develops. It likewise remains unclear what roles connectivity providers might play. AWS Local Zone, for example, is an edge computing service directly available to enterprises. Likewise, AWS Outposts puts AWS servers directly into an enterprise data center, creating yet another way AWS becomes a supplier of edge computing services.

Global Telco Share Prices Flat, Margins Dropping, But Not a Bad Outcome, Really

We should not be too surprised to learn that tier-one telco equity prices have not changed very much over the past few years. Companies in an industry with relatively slow incremental revenue growth, facing significant profit challenges, should find those trends reflected in their share prices. 

source: EY

Over a longer period of time, revenue has grown from the global recession of 2008 levels, but profit margins have declined. Some might predict a further dip as capital investment in 5G networks is made. Some of us think mobile operators have new tools and incentive to control such investments, in many cases even achieving 5G upgrades within the scope of existing capex budgets.

source: EY

The bottom line is that the telecom business, in monopoly or present competitive conditions, has never been a “growth” business, so we should never be surprised about relatively muted revenue growth or share price performance. 

The real battles are about maintaining modest growth, profit margins and cash flow as changes in the product mix and competitive challenges remain. It is not precisely “running in place,” more like “slow, steady ascent,” but that is a fairly reasonable way to describe a positive outcome. 

If outcomes are a bit boring, the execution challenges will be anything but.

5G Will Not Affect Consumer Mobile Revenues Much

There is a simple reason why it is unrealistic to expect 5G to alter retail connectivity provider revenue prospects very much: consumer ability to spend on communications is limited. In developed countries, for example, consumer spending on telecommunications services is about 1.5 percent to two percent of gross domestic product, and arguably is decreasing. 

source: ETNO

To be sure, spending can grow as GDP grows, but GDP growth is generally slow in developed markets, though faster in developing markets.

Also, while some other markets can grow because they become substitutes for existing products, communications actually faces the opposite problem: its products face substitution by lower-cost or free products. 

Yes, 5G will displace 4G accounts, but that is precisely the point: nearly every consumer 5G account simply replaces an existing 4G subscription. 

Some might therefore question the value of 5G. That misses a key point: the key immediate value is “more bandwidth at lower costs.” So 5G is important initially for service providers.

The eventual advantage is new enterprise applications enabled by 5G, in conjunction with edge computing, artificial intelligence and the internet of things. 

But on the consumer front, 5G is unlikely to move the revenue too much.

AWS Makes Bid for the Edge

It never has been clear how the edge computing market--particularly the "edge computing as a service" portion of that market, would develop. Plausible suppliers include cell tower companies, mobile operators, the hyperscale data center providers, data center providers or new specialists in the market.

But AWS is making big moves.

Amazon Web Services Inc. (AWS) has announced AWS Wavelength,  which provides developers the ability to build applications that serve end-users with single-digit millisecond latencies over a 5G network. 

Wavelength embeds AWS compute and storage services at the edge of telecommunications providers’ 5G networks, enabling developers to serve use-cases that require ultra-low latency like machine learning inference at the edge, autonomous industrial equipment, smart cars and cities, Internet of Things (IoT), and Augmented and Virtual Reality. 

AWS customers can now use the same familiar AWS APIs, tools, and functionality they use today, to deliver-low latency applications at the edge of the 5G network, around the world, AWS says. 

AWS is partnering with Verizon, Vodafone, SK Telecom, and KDDI) to launch AWS Wavelength across Europe, South Korea, and Japan in 2020, with more global partners “coming soon,” AWS says. 

It was probably inevitable that Verizon is partnering with Amazon Web Services for edge computing after AT&T and Microsoft announced their partnership for 5G-enabled edge computing.

Though each telecom giant is pursuing edge computing, the choice of a different strategic partner is a pattern with a long history, as doing so creates some amount of differentiation in the offers each can create. 

Verizon, says Amazon,  is the first technology company in the world to offer 5G network edge computing, and will use AWS’s new service, AWS Wavelength, to provide developers the ability to deploy applications that require ultra-low latency to mobile devices using 5G. 

The companies say they are currently piloting AWS Wavelength on Verizon’s edge compute platform, 5G Edge, in Chicago for a select group of customers, including video game publisher Bethesda Softworks and the National Football League. 

AWS and Verizon say the partnership brings “processing power and storage physically closer to 5G mobile users and wireless devices, and enable developers to build applications that can deliver enhanced user experiences like near real-time analytics for instant decision-making, immersive game streaming, and automated robotic systems in manufacturing facilities.”

AT&T calls its platform “Network Edge Compute,” which weaves Microsoft Azure cloud services into AT&T network edge locations closer to customers. That means AT&T customers will be able to use Azure services at the edge.

Enterprise Telephony Becomes a Feature of Productivity Suites-

By 2023, 40 percent of new enterprise telephony purchases will be based on a cloud office suite,  either Microsoft Office 365 or Google G Suite, analysts at Gartner predict. In part, that shift is the result of a couple of decades of declining investment in premises solutions and increasing investment in cloud-based solutions.

Some might say this change has voice communications becoming a feature of a productivity suite, in somewhat the same way that voice and messaging are features of a mobile phone's capabilities.

“As vendor investment in UCaaS has ramped up (while premises-based UC investment has been curtailed), UCaaS functionality now exceeds premises-based UC,” Gartner argues.

In fact, the old “private branch exchange” or business phone system market has been subsumed into unified communications, which includes traditional voice and telephony (fixed, mobile, private branch exchange, softphone), meeting solutions such as video conferencing or audio conferencing, voice mail, instant messaging and communications clients.

Much of What Matters in Life is Pretty Simple

How much of what matters in life is simply this?

How Much Share Does 5G Fixed Wireless Have to Shift to be Meaningful?

How much market share does 5G fixed wireless have to shift before it affects the profitability of the fixed network consumer internet access market? Not much. 

In recent quarters, for example, U.S. fixed network internet access net additions have totaled about six tenths of one percent of the installed base, with cable gaining eight tenths of one percent while telcos lost about two tenths of one percent. 

In other words, a shift of about two-tenths of one percent per quarter halts the telco decline. A shift of perhaps six-tenths of one percent--from cable to telco--actually causes cable share to begin a decline. 

That is what the stakes realistically are: a chance for telcos to halt, and perhaps reverse, the long-term decline of their market share in internet access. 

Cable TV executives in the U.S. market naturally express as much skepticism about the dangers 5G fixed wireless services pose for their consumer broadband business as telco execs say they are optimistic. Basically, it will no scale, or will scale too slowly to keep up with cable’s own planned bandwidth plans, cable execs tend to say. 

There is reason for the cable views. The threat of optical fiber to the home has existed for a few decades, but has not dented cable’s emergence as the leading supplier of consumer internet access connections using fixed networks. Cable has about 67 percent of the installed base and has essentially gotten more than 100 percent of the net new account additions for a couple decades. 

In fact, over the last 20 years, it would be hard to find a single year where cable broadband net account gains were not about 60 percent to 70 percent of all net gains. 

source: NTIA 

source: Broadband Trends 

source: Leichtman Research Group

But the impact of 5G fixed wireless is not that it dramatically upsets the market. 5G fixed wireless might be the only way telcos collectively can halt a long-term decline of their market share. 

It might be prudent not to envision any scenario where 5G fixed wireless actually upends the market share structure. Instead, 5G fixed wireless probably is relevant because it might shift just enough share to choke off the cable growth model, reverse the telco share loss trend, and then change profit margins.

It is very subtle stuff. Verizon, for example, only has to gain about 7,000 5G fixed wireless accounts per quarter to halt its customer losses. T-Mobile US and Sprint have virtually zero fixed network market share, so even smallish gains represent new accounts with average revenue possibly double what they get from mobile internet access accounts.