Communications "Just an API?"

“Bits are bits,” many used to remark during the early days of transition from single-purpose to multi-purpose networks. 

“Communications” is “just an application programming interface," some now argue.

Both statements are true to an extent, and wrong to an extent. It is true that the advantage of multi-purpose networks is that they can carry any media type. In that sense, bits are bits. But various applications require different support. 

Video requires consistent packet delivery and hence created a need for content delivery networks, which reduce latency and jitter artifacts that mar video experience. 

source: Global Dots 

Various forms of extended reality or process control  will likely require intensive data processing very close to the end user, hence creating a need for edge computing. 

Likewise, what tends to be meant by “communications is just an API” is that communications capabilities (voice, messaging, unified communications) can be added to any application using an API. 

It is meant to mean the abstraction of a great amount of complexity and infrastructure that the actual end user need not worry about. 

Comcast Boosts Speeds, Adding 3 Gpbs Symmetrical Service

Comcast has increased speeds for its most popular Xfinity Internet tiers for residential customers across 14 northeastern states from Maine through Virginia and the District of Columbia.

In October 2021 Xfinity increased Gigabit Pro speeds from 2/2 Gbps to 3/3 Gbps symmetrical nationwide. That service is offered using a fiber-to-home connection. 

source: Comcast

Also, the 800 Mbps tier will have speeds boosted to 900 Mbps, with an additional 5 Mbps added in the upstream. The 400 Mbps tier will be boosted to 600 Mbps, with a 10 Mbps increase in upstream speeds.

The 200 Mbps tier will be upped to 300 Mbps, with 5 Mbps faster upstream speeds.

Home Broadband Prices Dropped 10% in 2021, Speeds Increased 32%

If there are any iron rules about the home broadband business, it is that prices tend to fall while the amount of supplied bandwidth tends to increase. 

In 2021, the average monthly cost of fiber-based residential broadband decreased nearly 10 percent, according to Point Topic.

In the same period, the price of cable broadband increased by 1.2 percent,  while the average subscription to copper-based broadband services increased by 0.5 percent.

The average monthly charge for residential broadband services fell by 10.4 percent in 2021 and cost $82 per month, adjusted using the purchasing power parity method.

At the same time, the average downstream bandwidth provided to residential subscribers has gone up by 32 percent.

Global Bandwidth and Retail Price (adjusted for purchasing power parity)
Region (Residential Broadband)	Average Downstream Speed, Mbps (Q4 2020)	Average Downstream Speed, Mbps (Q4 2021)	Average Monthly Charge, $PPP (Q4 2020)	Average Monthly Charge, $PPP (Q4 2021)
Asia-Pacific	856	1,355	$    63.45	$ 61.53
Eastern Europe	232	278	$    66.68	$ 66.46
Latin America	158	254	$  166.38	$  115.62
Middle East/ Africa	91	155	$  160.93	$  155.28
North America	358	396	$    83.93	$ 88.31
South, East Asia	371	393	$    83.16	$ 66.43
Western Europe	325	403	$    68.00	$ 68.20

source: Point Topic

NTT Docomo Tests Hybrid Cloud with AWS as Part of Move to Cloud-Native 5G Core Network

NTT Docomo and NEC Corp. have launched  proof-of-concept (PoC) testing to run NEC's 5G core network service in a hybrid cloud environment that leverages Amazon Web Services for cloud-native mobile networks. 

source: NTT

As the first step, 5GC network functions will run in the cloud on AWS infrastructure and corresponding 5GC network functions will reside on DOCOMO's on-premises infrastructure to verify availability and operational feasibility, says NTT. 

The second step will deploy the 5GC network function on AWS Outposts (on Amazon EC2 compute instances powered by Graviton2 processors) on DOCOMO's premises. 

In an era where fundamental computing architecture has shifted to the cloud (remote computing), and where use of public cloud resources follows, the test simply makes sense. Where “computing” once happened primarily on closed and specialized mainframes (Class 5 central office switches), most computing now happens on servers that are increasingly distributed. 

Some will worry that by “outsourcing” computing platforms, connectivity service providers are in danger of lessening their value in the ecosystem. 

But control of computing processes is one matter; how the computing platform is sourced is a different matter. Every enterprise uses a specific mix of resources and apps to support its core business processes. 

At least in principle, it matters little--except for cost reasons--whether the hardware is owned or leased; how core software is purchased or where the processors reside. 

Telcos and other connectivity providers long ago gave up the design and manufacture of their core hardware and software resources to third parties, just as they mostly gave up the design and manufacture of end user devices. 

The shift to public cloud for computing resources is simply the latest shift.

Decoding "Network as a Service:" On-Demand

Some of you might find the network as a service “strategy” for connectivity providers to be a bit baffling. Connectivity has always been a service, with a few salient exceptions, all centered around private and local networks. 

Local area networks have mostly been a “do it yourself product” where organizations or individuals own their own infrastructure. Wi-Fi has operated mostly the same way. Enterprise voice services often have used phone systems (private branch exchanges) that use owned hardware and software to “create voice services” similar to--but extending beyond--those provided by the public voice networks. 

The shift to “software as a service” and “computing as a service” are the better examples of former products that now are sold as services. Consider almost anything Cisco might sell, offered as a subscription, without ownership, as a better example. 

“NaaS is a cloud model that enables users to easily operate the network and achieve the outcomes they expect from it without owning, building, or maintaining their own infrastructure,” says Cisco. 

In the connectivity services business, that “as a service” framework already exists. Network as a service makes complete sense in the information technology infrastructure business. One does not buy computers, one leases compute cycles. 

As Verizon uses the term, network as a service primarily refers to any services Verizon can deliver on demand, customized or personalized. 

source: Verizon   

 

So how are “connectivity services” understood by the shift of terminology to “network as a service?” 

source: Verizon

Are Verizon and T-Mobile too Optimistic About Fixed Wireless, or are Skeptics too Pessimistic?

It appears that Verizon might be both "too optimistic" and also "not optimistic enough" about fixed wireless account growth, all at the same time. In either case, pessimists might be wrong.

If present trends continue, Verizon might be understating the importance of fixed wireless as a means of gaining home broadband share. In the fourth quarter of 2021, 38 percent of all home broadband net account growth was gained by fixed wireless networks.

It is clear enough that there is much skepticism and hope about about the impact of 5G fixed wireless. T-Mobile and Verizon believe they will be significant winners in terms of taking home broadband share. 

Verizon, in fact, believes fixed wireless will provide the initial incremental revenue upside from 5G, generating more additional cash than longer-term opportunities such as edge computing. Nobody would accuse T-Mobile of being too pessimistic, however.

T-Mobile expects to have between seven million to eight million fixed wireless accounts by 2025, and views an addressable market of about 30 million homes that are suitable from a signal quality and capacity standpoint. 

Craig Moffett, analyst with MoffettNathanson, views T-Mobile's forecast, which implies 23 percent to 30 percent penetration of addressable homes, as an "arguably absurdly ambitious target."

But T-Mobile could be proven correct, and not chasing an overly-ambitious goal, to a point. If one assumes T-Mobile gets a disproportionate share of its fixed wireless accounts in rural areas, and if one assumes fixed network service providers in those rural areas will not, by 2025, significantly upgraded networks to either fiber to home or DOCSIS 3.1 standards, then T-Mobile could well grab 20 percent to 30 percent market share.

Perhaps the more-germane criticism is what happens in urban and suburban areas, longer term. Eventually, as “typical” home broadband users consume more data, fixed wireless capacity will have to increase as well.

To be sure, there are ways to compete. Millimeter wave assets are an obvious solution. More mid-band spectrum helps as well.

Small cells will help boost usable bandwidth by increasing specrum reuse. And usage patterns will play a role as well. Some argue that peak home broadband will happen at night, when usage of the mobile network is lowest.

Still, there are obvious challenges, if fixed wireless demand scales.

OpenVault data shows that the average U.S. cable broadband household already uses about 434 gigabytes of data per month, roughly 40 times more than the average unlimited usage plan mobile customer. 

So if new fixed wireless households impose 434 GB of demand on the wireless network, how many customers can be added before there are capacity constraints? And how well will the coping mechanisms work?

Also, will T-Mobile and Verizon fixed wireless using mid-band assets primarily be attractive to "average" users or the "lighter user" segment? And will Verizon's millimeter wave network be extensive enough to reach further up the scale of usage support?

Verizon, meanwhile, is committing to about $1 billion in fixed wireless revenues by 2024, which MoffettNathanson believes implies 1.7 million customers. Verizon both agrees--and appears to disagree.

Verizon suggests it will add 4.5 million net new home broadband accounts by 2025. It also says fixed wireless will represent “more than one million” of those accounts. 

source: Verizon

The numbers seem incongruous. Fios already has adoption in the 40-percent range. So assume Verizon can boost adoption to 50 percent. That would only add about 650,000 additional Verizon Fios home broadband accounts. The rest has to be fixed wireless or acquired through purchases of existing accounts. The latter is unlikely to happen, so the growth has to be organic. 

Nearly all of those accounts would have to be generated by the “out of region” fixed wireless network, to make the numbers work. To be sure, perhaps Verizon is banking on adding millions of new Fios-passed locations. 

Verizon says it passes 16.5 million locations with Fios at the end of 2021, and expects to pass 18 million locations by the end of 2025. So add 1.5 million new Fios locations. Assume adoption is at 50 percent. That suggests a potential 750,000 accounts. 

All together, that suggests 1.4 million new Fios accounts by 2025, with total net gains of 4.5 million accounts. That implies something more than three million fixed wireless accounts gained by 2025. 

source: Verizon

 

source: Verizon

The point is that either Verizon seems conflicted about the size of the opportunity, to some extent. T-Mobile is viewed as too optimistic by some. But fourth quarter 2011 share gains by fixed wireless were unprecedented.

Fixed Wireless Had Monster 38% Share of Net New Account Adds in 4Q 2021

In an unprecedented occurrence, fixed wireless services provided by T-Mobile and Verizon alone represented 38 percent share of broadband net adds in the fourth quarter of 2021, according to MoffettNathanson. 

All skepticism aside, that has never happened before in the U.S. broadband market. So while we might still disagree about the size of the fixed wireless opportunity, early returns suggest 5G fixed wireless might have significant upside. 

source: LightReading, MoffettNathanson 

In fact, were it not for fixed wireless net additions, telcos would have lost market share in the fourth quarter of 2021. 

Concerns about limited bandwidth remain, but consumer adoption trends suggest those concerns are possibly overblown. At least for the moment, consumers see enough value to switch. 

About half of Verizon's fixed wireless customers are coming from commercial accounts, MoffettNathanson says. 

T-Mobile says nearly half of its fixed wireless customers represent market share gains at cable operator expense, the firm says. 

Sector	Q4 2021 Gain/Loss	Q4 2020 Gain/Loss	Year-on-Year Growth %	Total
Cable	+464,000	+899,000	+3.8%	79.43 million
Telco	-26,000	+21,000	-0.4%	33.51 million
FWA*	+302,000	+81,000	+463.9%	869,000
Satellite	-35,000	-35,000	-6.6%	1.66 million
Total Wireline	+437,000	+920,000	+2.8%	112.95 million
Total Broadband	+704,000	+966,000	+3.3%	115.48 million
* Verizon and T-Mobile only

source: LightReading, MoffettNathanson