Telcos Running 5G Will be Big Users of Their Own Edge Computing Networks

Edge computing is not, strictly speaking, “new.” Content delivery networks have for decades provided edge acceleration services for content and app providers, proving lower-latency app performance. And that might remain true for much of the next wave of edge computing investments as well. 

Omdia analysts expect a doubling of edge deployments through 2024, involving new 4.7 million servers. Enterprises, hyperscale computing as a service suppliers and communications service providers are among the industry verticals with significant deployment plans. 

You might not think telcos would be such early adopters, but many own streaming video services, which makes such telcos clients for content delivery. By 2024, telcos will be among the biggest users of edge computing, for that reason, according to Omdia. 

source: Omdia

Also, virtualized networks require more compute on the edge of the telco network, to support end-to-end virtualized network functions, which distribute core network functionality out to the edges of the transport network. 

source: The Fast Mode

As network services shift inexorably to the edge, Omdia predicts edge server workloads will expand to include self-driving car telemetry, augmented and virtual reality (AR/VR) applications and cinema-quality gaming. But some of those apps may take time to develop. 

The widespread deployment of fully automated driving systems that have no safety driver onboard will take at least a decade, a new study by MIT suggests. 

source: Wipro

Business Travel Will Escalate Once Firms Start Losing Accounts

The conventional wisdom in many quarters is that "nothing will be the same" post-Covid. It is counter-intuitive, but I am betting that not so much change will be permanent, beyond those trends already happening before the pandemic. 



That is not to say "no changes" will be permanent, simply that the magnitude will likely not be as great as some believe. Consider business travel, where many now believe "never" will return to former levels. Yes, there will be reluctance, for a while. 



But once competitors start taking accounts, businesses will do what they believe they must do to defend market share. Working from home will remain popular among employees until promotions start to favor those who are "in the office." 



Up to this point, with lockdowns, every company has been on a level playing field. In the short term, productivity slippage has not been seen. Whether than remains true longer term is the question. Will some firms benefit more than others from informal interactions at work?



That always is said to be advantages of employees bumping into others they do not typically work with. After the pandemic, when firms will have choices, how will perceived advantages be seized? And how much "life as before" might play a role in providing advantages?

American Tower Activates New Edge Computing Center

Collocation is among the potential roles and revenue streams for edge computing, with cell tower owners, telcos, retailers, modular building and others seeking roles.  Colo Atl, a subsidiary of American Tower, for example, has activated a new edge data center in Atlanta, located in the Ben Hill, GA area, approximately 10 miles southwest of downtown Atlanta.

American Tower Edge offers edge data centers at the base of cell towers in six cities including Atlanta, Jacksonville, Denver, Boulder, Austin and Pittsburgh.

By some estimates, most of the link latency happens in the access network from device to cell tower, so placing a data center at the base of a tower eliminates the greatest source of link latency between an end user device and a computing resource. 

source: Vertical Bridge

The issue is the business model, as it might not make sense to put an edge computing center at the base of every cell tower. Many believe siting an edge data center somewhere “within a metro” is generally applicable, as ultra-low-latency apps will be supported by premises edge computing or “onboard the device” computing. 

An interconnection and colocation facility, the new data center is said to offer customers an alternative location to the centralized metro data centers in the metro area.

he purpose-built 360-square-foot, 100 KW facility provides eight customer cabinets, consisting of twenty quarter-cabinet lockers and three full cabinets. The connectivity enabled, multitenant, neutral-host edge data center maximizes flexibility and is directly connected via dark fiber to Colo Atl, the company says. 

SBA Communications operates SBA Edge, while Crown Castle is partnering with edge specialist Vapor IO to build a national network of edge computing sites. 

AI is Computation and Memory Intensive

Machine learning is computationally and memory intensive. AI capabilities, for example, have increased by a factor of 300,000 from 2012 to 2019, a doubling every 3.43 months, says Rambus. 

source: Rambus

Memory requirements likewise will grow with use of AI. Advanced driver-assistance systems will require memory bandwidths exceeding 200 GB/s, says Rambus. 

source: Rambus

Is Edge Computing a Substitute for Network Slicing?

It would be fair to note that connectivity providers often have developed and offered  many features and services they  believed customers would want. The catch is that revenue for connectivity providers is cost for enterprises, businesses and consumers who are asked to buy and use the features and services. 

Connectivity providers tend to prefer selling turned-up services; enterprises often prefer to buy dark fiber. Sometimes regulations prevent sales of enhanced quality products, even if there is conceivable consumer demand. Best effort internet access provides an example. 

It is conceivable there is demand for latency and jitter-controlled internet access, but regulations bar the sale of such products under the rubric of network neutrality.

source: TM Forum

Network slicing, the ability to create virtual private networks that run end to end on 5G networks, provides another opportunity to find out where demand might lie for private networks whose characteristics and performance are better matched to some use cases. 

Of course, to the extent network slicing offers business value, potential buyers will have incentives to explore “do it yourself” alternatives when it saves them money. In that sense, edge computing networks are an alternative to network slices. 

source: TM Forum

If ultra-low-latency applications are those which could benefit from network slices, one alternative is do commuting at the edge, and not sending data across wide area networks that are optimized for low latency. In many use cases, the value of ultra-low-latency computing is supplied by edge computing services, with non-real time backup across wide area networks. 

Perhaps ironically, consumer customers who have few other alternatives might be good candidates for internet access with quality of service features a network slice offered by a connectivity provider. But regulations often prevent such offers. Gaming services, work from home conferencing and ultra-high-definition video are among potential use cases. 

What is edge computing?

IBM, Verizon Business Collaborate for MEC

A new collaboration between Verizon Business and IBM illustrates the way 5G, multi-cloud support, edge computing, artificial intelligence, internet of things, “Industry 4.0,”  private networking are intrinsically related. 

The companies plan to combine Verizon’s 5G and Multi-access Edge Compute (MEC) capabilities, IoT devices and sensors at the edge, and IBM’s expertise in AI, hybrid multi cloud, edge computing, asset management and connected operations. 

source: IBM

The collaboration uses Verizon’s ThingSpace IoT Platform and Critical Asset Sensor solution (CAS) plus IBM’s Maximo Monitor with IBM Watson and advanced analytics. This effort has IBM supplying the needed analytics and multi-cloud computing support; Verizon the edge devices, access network and collocation facilities. 

IBM and Verizon are also working on potential combined solutions for 5G and MEC-enabled use cases such as near real-time cognitive automation for the industrial environment. 

Separately, Verizon says the 5G Future Forum will release its first technical specifications in the third quarter of 2020. The 5G Future Forum is a group of 5G service providers and suppliers working to  accelerate the delivery of Multi-access Edge Computing-enabled solutions around the world.

The 5G Future Forum was established in January 2020 by América Móvil, KT Corp., Rogers, Telstra, Verizon, and Vodafone.

Webscale and Connectivity Providers in Multiaccess Edge Computing: Who Wins?

Executives discuss prospects for multi-access edge computing for hyperscale computing as a service providers and connectivity providers. It still is not clear how the segments will benefit, though the argument is made there is room for both webscale computing service providers and connectivity providers. 

Indranil Chatterjeee SVP of Products, Sales & Marketing
Dr. Alex Jinsung Choi SVP Strategy & Technology Innovation	Glenn Gore Chief Architect
Dennis Lauwers Distinguished Engineer, Hybrid Cloud Europe

Edge Computing Total Cost of Ownership: Volume Matters

Generally speaking, enterprise do it yourself information technology makes business sense when volumes are high, as it generally makes sense to own infrastructure used at high levels, rather than renting capacity. 

At low volumes of use, renting often makes more sense. That seems to apply for enterprises looking at edge computing, according to a Mobile Experts analysis. Over three years, “it costs about 35 percent to 55 percent more to access a hyperscaler's cloud for a heavy industrial workload at the edge,” Mobile Experts says.

That mirrors what we tend to find for other information or communications choices as well. At low volumes, it almost always makes sense to use a hosted “voice as a service” solution. At high volume (at large headquarters locations, for example) it virtually always makes sense to buy and operate voice servers. 

At least in part, that is because voice as a service is purchased “by the line,” while owned servers, up to a point, impose no additional recurring costs for heavy usage. 

cloud storage veruss on premises storage shows similar total cost of ownership curves. In many cases, in high volume, premises storage has lower total cost of ownership. Many enterprises find the same trend for computing cycles: in lower volumes, cloud can be cheaper; at high volumes a premises solution might be cheaper. 

5G Operators will be Core Users of Network Slicing and Edge Computing

Network slicing, the ability of a 5G core network to create virtual private networks, also is related to edge computing, since network slicing moves user plane functions (apps and use cases) towards the edge of the core network. 

That applies to the core 5G network itself, as much as any retail slices sold to customers, since the virtualized network itself uses user plane control at the network edge. So one important use case for network slicing is the core operation of the 5G network itself. 

As many hope, the ability to create such virtual networks, optimized for different use cases, also will allow greater opportunities for customized use cases with revenue implications for mobile operators using 5G virtualized core networks. 

source: Whatis

Network slicing therefore has both an internal use case--supporting the operation of the core 5G network itself--as well as third party customer use cases. And many of the expected new use cases are intended for use in business-to-business settings, not consumer use cases, especially where ultra-low latency performance is essential or quite important. 

source: Researchgate

For example, a mobile operator will use a slice of its own to provision core voice services, messaging, consumer and business internet access. Other slices could be ordered up by private network users as well. 

source: IEEE

Edge Computing and Small Cell Networks

Edge is Contested Terrain

As always, connectivity providers provide the highest value when “connecting” servers and other servers, people or machines. In relationship to edge computing, some of the traditional roles likely will prevail. 

Connectivity providers terminate their networks “at the side of the building,” with the consumer or business operating their own private premises networks (local area networks).

source: AWS

So wide area network transport and local access to the site are the places where connectivity providers traditionally make their money. Edge computing is contested terrain. The function still is “computing,” so the cloud hyperscalers have a role, essentially extending computing with ultra-low-latency requirements closer to the places data has to be processed. 

source: AWS 

But many other contestants hope to benefit as well, providing colocation facilities, edge computing server infrastructure, edge apps, integration services or actual “edge computing as a service” offerings. 

It remains to be seen which firms ultimately emerge as leaders in those respects. Some service providers hope to become “edge computing as a service” providers. Others see partnerships withy computing as a service suppliers as more logical, where colocation and facilities are the primary value supplied, along with connectivity in the local access area.

Edge Computing is Part of 5G Plans in Asia, Pacific

Telstra, VHA, Bharti Airtel, Reliance Jio, Vodafone IDEA, Rakuten, SK Telecom, KT, China Mobile, China Unicom, and China Telecom are among telcos who believe edge computing is a sizable revenue opportunity.

Telco edge computing in those cases will develop in parallel with 5G Standalone deployments starting in 2021, says International Data Corp. and often will take the form of virtualized Infrastructure-as-a-Service (IaaS), with the business model based on a multi-tenant distributed edge cloud ecosystem. 

In other words, telcos hope they can supply edge computing as cloud hyperscalers supply cloud computing services, especially when ultra-low-latency and high-bandwidth use cases must be supported. 

Autonomous transportation, vehicle communications and computing, artificial reality, virtual reality, high-performance gaming, and real-time sensory and image processing provide examples of possible use cases. 

IDC also believes network slicing to create virtual networks for other web-scale companies such as Microsoft, Google, AWS, Tencent, Baidu, Alibaba, and their ecosystem partners likewise will be a new revenue opportunity. On-demand, infrastructure as a service and wholesale are possible pricing options.

Who Will Own the Edge?

“Who owns the edge?” is a very good question for all would-be participants to be asking. So far, a few general categories of use cases seem universally cited by standards groups, service providers and would-be infrastructure or service suppliers, all playing to the new capabilities of 5G networks. The 3GPP framework is built on capacity (“enhanced mobile broadband”), massive internet of things and ultra-reliable low-latency communications. 

Everyone will notice that the value proposition is all about ultra-low latency and capacity, for business-to-business and consumer apps. According to Vertiv, that means use cases including security, smart grids, smarter retail, augmented reality and haptic computing, virtual reality,  smart cities and life-critical use cases. 

source: Vertiv

Virtually nobody would disagree that all those use cases are conceivable uses for networks and computing architectures with so-far unparalleled latency performance. 

But that also means the edge looms large because ultra-low latency, while fine for the access connection, does not offer as much value if compute functions do not match the access latency. 

Roles and revenue opportunities are the issues for connectivity providers. Presumably there are connectivity account opportunities, in the form of new mobile or fixed connections, higher customer spending on capacity or colocation for partners who want to put edge computing facilities into place. 

Less certain are opportunities in the actual edge “computing as a service” part of the business. . So far, despite the rhetoric, most tier-one telcos with edge computing aspirations still are focusing on connectivity revenues or real estate (colocation in telco facilities, for example). 

That does not mean other opportunities will fail to develop, perhaps in the way that Verizon acquired Bluejeans, the videoconferencing service. There is ample precedent for telcos acquiring apps, content services, media platforms, connected car and other “services” that rely on communication networks. 

Still, we remain in early days.