Business White Paper - Telco Edge Computing

Executive Briefing Service
Kindly supported by:
Executive Briefing
TELCO EDGE COMPUTING:

WHICH USE CASES AND HOW
TO MONETISE?
Having spent years evaluating the edge computing concept and potential
opportunity, telcos are now in a position to take action. Many have
started to make decisions on where the edge is and how it will be built.
This report focuses on how telcos can monetise their edge, deep diving
on gaming, video analytics, predictive maintenance and healthcare.
Dalia Adib, Practice Lead – Edge Computing | dalia.adib@stlpartners.com | October 2020

TELCO EDGE COMPUTING: WHICH USE CASES AND HOW TO MONETISE? | OCTOBER 2020
Executive Summary
Telcos should consider 5 key principles to monetise edge computing
At a high-level, the benefits of edge computing are broadly well-understood. There is an acceptance
that certain applications will benefit from lower latency, reducing the amount of data that needs to be
sent over networks (and associated costs), complying with data privacy more easily, amongst other
benefits. Progress has also been made in developing a common way of characterising different types
of edge: from on-premises to the network edge.
However, telcos are still grappling with developing commercial models that will allow them to
monetise edge computing and create value using their existing assets and capabilities, as well as new
ones. One of the challenges is that the edge market needs nurturing – telcos need to work with their
customers and the wider ecosystem to seed the demand for edge computing. This requires driving
engagement between those providing edge infrastructure (from facilities to software), networks, end-
devices and sensors, applications, and then orchestrating this altogether. As telcos scale the business,
orchestration will be key to managing customer workloads and the underlying infrastructure to best
make use of finite edge resources while ensuring a consistent quality of service regardless of the use
case.
One of the ways in which telcos can kickstart their monetisation strategy for edge, without necessarily
having the end-to-end business case ready from Day 1, is by adhering to five key principles and taking
an iterative approach.
Figure 1: 5 principles for commercialising edge computing
1. LEARN
5. SCALE edge
the business benefits by
effectively industry
4. CO-CREATE
2. SELECT
with the
ecosystem use cases
3. DEFINE
business
models
Source: STL Partners
© STL Partners EXECUTIVE BRIEFING 2

There are many use cases that will benefit from edge – telcos can
explore these as discrete use cases or vertical domains
For telcos, the number of potential use cases can be overwhelming, but it is important to focus on a
small number of use cases initially and begin to engage customers within specific industries, building
a solid minimum viable proposition that is relevant to at least some customers before developing new
services.
These can be considered discrete use cases that may span multiple industries. Some of the prominent
use cases being evaluated by telcos are highlighted below. Telcos might also explore the potential of
addressing vertical sector domains with a variety of use cases, for example smart healthcare. This
allows telcos to focus on building expertise in a single sector and streamline investments into
application platforms and services for that industry too.
Figure 2: Example use cases evaluated by edge value and use case maturity
Use cases Value of edge Use case maturity

AR/VR field assembly and safety
Connected car driver assistance
Edge ADN/CDN
Low-latency cloud gaming
Private cellular networks
Advanced predictive maintenance
Security – video analytics
This report deep dives on three use cases and one vertical domain:
• Low-latency cloud gaming
• Security video analytics
• Advanced predictive maintenance
• Smart healthcare
Next steps for telcos: putting a monetisation strategy into action

The key recommendation in this report is for telcos to start executing a practical plan for
commercialising edge computing by following the five key principles outlined. Applying these at a use

case level (as shown below) can help to bring telcos’ value propositions to life and develop a business
case, at least, for an individual industry or use case.
In order to do this effectively and engage customers and partners to take services to market, telcos
should consider how they can build common frameworks, forums and platforms to onboard
ecosystem partners, co-innovate, conduct proofs-of-concept and develop commercial models. Being
able to orchestrate the ecosystem, the underlying edge infrastructure and services efficiently will
ensure success in monetising edge computing.
Figure 3: Five principles applied to use cases

1. Learn edge 2. Select use 3. Define 4. Co-create 5. Scale the
Use case
benefits by cases business with the business
domain
industry models ecosystem effectively
Reduce latency/ping/ Related use cases Applications: gamers Content is king; Consider the
lag could include IaaS/PaaS: gaming studios, aggregators orchestration
Reduce end-device matchmaking platforms and developers with capabilities required
Low-latency TCO players, AR/VR Dedicated edge the best games are to provide edge-
cloud gaming Enable new types of gaming hosting: cloud critical enabled cloud
games and models providers Develop potential gaming across
platforms / APIs with multiple devices and
games engines networks
Reduce backhaul and Related use cases Dedicated edge Work with ISVs that Select one industry
TCO for customer could include hosting or IaaS/PaaS: are specific to each segment to work with
Enable new business condition-based Industrial OEMs/SIs industry segment and to start with and build
Advanced models for monitoring and ISVs SIs to integrate into MVP with sensor
predictive maintenance Solutions: industrial CRM / ERP systems, manufacturers and
maintenance Gain customer end-customers leverage existing IoT ISVs, leverage
insights relationships where existing IoT
possible platforms, define role
of network edge
Reduce backhaul Sub-use cases Dedicated edge Start with single Start with single
Allow customer to include real-time hosting or IaaS/PaaS: security use case and security use case and
access innovative object/event Security SIs and ISVs develop MVP from develop MVP from
Security – applications without identification, health Solutions: enterprise there, then work with there, then work with
video changing and safety end-customers customers to customers to
analytics infrastructure understand how this understand how this
can be extended to can be extended to
other video analytics other video analytics
applications applications
Access to cloud Sub-use cases IaaS/PaaS: IoT / Need to work with Ensure access to
services without include: remote application providers established remote patient
using public cloud patient monitoring, in- Dedicated edge healthcare specialist monitoring
Enable remote patient patient monitoring, hosting: SIs or private SIs and OEMs that applications by
Smart electronic healthcare healthcare providers orchestrating
monitoring cloud providers
healthcare records trust and leverage workloads across
Maintain data privacy Applications /
solutions: consumer local (public sector) multiple edge sites
or healthcare relationships within for ubiquitous
providers market population coverage

Table of Contents
Executive Summary ............................................................................................................................................... 2
Introduction to telco edge computing ................................................................................................................ 7
Key principles for telcos seeking to monetise edge computing ................................................................... 9

1. Learn the edge benefits for each industry ................................................................................................. 9
2. Select use cases .......................................................................................................................................... 10
3. Define the business model ......................................................................................................................... 11
4. Co-create with the ecosystem ................................................................................................................... 12
5. Scale the business effectively ................................................................................................................... 13
Use case deep-dive .............................................................................................................................................. 14

Low-latency cloud gaming.............................................................................................................................. 14
Video analytics – security ............................................................................................................................... 16
Advanced predictive maintenance ................................................................................................................ 17
Smart healthcare .............................................................................................................................................. 19
Next steps.............................................................................................................................................................. 22

Table of Figures
Figure 1: 5 principles for commercialising edge computing.......................................................................... 2
Figure 2: Example use cases evaluated by edge value and use case maturity .......................................... 3
Figure 3: Five principles applied to use cases .................................................................................................. 4
Figure 4: Telco edge computing locations ........................................................................................................ 7
Figure 5: Virtualised RAN architecture uses edge computing ....................................................................... 8
Figure 6: The edge computing value chain ....................................................................................................... 8
Figure 7: 5 principles for commercialising edge computing .......................................................................... 9
Figure 8: Benefits of edge computing .............................................................................................................. 10
Figure 9: Example use cases evaluated ........................................................................................................... 11
Figure 10: Edge computing business models ................................................................................................ 12
Figure 11: Each edge use case leverages different benefits and may use a different type of edge
location................................................................................................................................................................... 14
Figure 12: Potential partners in cloud gaming ................................................................................................ 16
Figure 13: The ecosystem required for advanced predictive maintenance .............................................. 18
Figure 14: Smart healthcare use cases ........................................................................................................... 20
Figure 15: Five principles applied to use cases .............................................................................................. 22

Introduction to telco edge computing

The concept of edge computing centres on putting flexible, IT infrastructure anywhere on the
spectrum between the cloud and the device. This is, in some cases, to bring the cloud closer to the
end user or the end device, rather than in a remote data centre. In other cases, the advantages come
from moving away from using highly “local” infrastructure, traditional on-premises (using proprietary
appliances and systems) or on the device, to an environment more akin to cloud and virtualisation.
Edge computing will be important to run the network in a more flexible, distributed manner by placing
virtual network infrastructure at sites deeper in the network. However, it also presents telcos with an
opportunity to create new services by leveraging that same edge computing infrastructure to run third
party applications.
One of the opportunities, and also challenges, for telcos is that they have a vast choice of different
types of edges available to them. Telecoms operators are putting edge computing at the network edge
– a near edge for them – by evolving sites such as central offices and base stations to host mini data
centres. Many telcos may also want to provide edge computing services to customer sites. For
example, putting an edge compute capability in enterprise customer premises equipment (CPE) to run
enterprise applications, or providing a small edge compute footprint at the manufacturing site, office
space, retail shop, etc.
Figure 4: Telco edge computing locations

Physical location of compute workloads or processing
Enterprise Radio Access Network Metro Core Network Internet
Telco edge computing
“Far edge” Network edge

“Near edge”
On-premises edge
Telco edge computing is not independent from other initiatives happening in the telco space and what
makes it exciting is that it can both enhance and leverage other technologies telcos are investing in
(5G, telco cloud, virtualised RAN) to change the business model and open up new revenue streams.
The table below summarises these:
Initiative Relevance to edge computing

• Many 5G applications require both 5G and edge computing to achieve round-trip low latency on a
consistent basis
5G
• Edge computing-enabled applications allow telcos to take 5G-like services to enterprises before there is
full coverage of 5G

• Edge compute infrastructure may be used to deploy private 5G networks at enterprise sites – as well as
be used to host applications on-site
• 3GPP has defined the enablers for multi-access edge computing (MEC) in its 5G standards
• The virtualisation of the network is resulting in the transformation of telecoms sites and infrastructure to
be more IT-like; telcos will be able to scale the infrastructure, skills and capabilities by not only using
them for internal network infrastructure, but also third party applications and MEC platforms
Telco
• Telco cloud is not only about virtualising network functions, but creating a platform for innovation, which
cloud
will allow telcos to leverage flexible network services to complement edge computing applications
• An extension of this is network slicing – combining compute and networking to create differentiated
services customised to a particular enterprise or software application
• Although vRAN is a subset of telco cloud, it is particularly interesting in the context of edge computing
Virtualised given that the physical location of a radio access network means that VNFs for the RAN will
RAN undoubtedly use edge computing equipment (see figure below)
(vRAN) • Sites that will host vRAN will be refurbished to become more data centre-like environments, making
them ready to host edge computing systems for third party applications and MEC platforms too
Figure 5: Virtualised RAN architecture uses edge computing

Base station
Antenna
Radio Unit
Virtual
Baseband Unit
Edge server
Core network
Internet
Despite various factors making it a ripe prospect, telcos cannot take the edge computing opportunity
as given. Edge computing use cases are still at a nascent stage and the ecosystem is still developing.
The telecoms industry needs to understand the needs of customers, define clear monetisation models
and work with these customers and other players in the value chain to succeed.
Figure 6: The edge computing value chain
Edge cloud Systems

Facilities Hardware infrastructure
Application Networks Orchestration Devices
integration

Key principles for telcos seeking to monetise

edge computing
Building edge computing infrastructure is no easy task: each type of edge requires a different
hardware and software stack due to the size of the deployment, and the variety of potential use
cases means that market requirements are not easy to consolidate.
Telcos need to ensure they have a clear plan to monetise edge computing to ensure they make
returns on their investment. This is not only about technology, but the way in which the telco
develops a commercial plan and meets the needs of customers – whether they be businesses,
consumers or technology partners, such as independent software vendors (ISVs) or systems
integrators (SIs).
We believe there are 5 key principles telcos should consider when evaluating the edge opportunity.
Figure 7: 5 principles for commercialising edge computing
1. LEARN
5. SCALE edge
the business benefits by
effectively industry
4. CO-CREATE
2. SELECT
with the
ecosystem use cases
3. DEFINE
business
models
1. Learn the edge benefits for each industry

Telcos need to understand why customers would even consider leveraging edge computing,
particularly compared to existing alternatives – whether it be cloud or traditional proprietary systems
using on-premises computing. Much of the discussion has been on the ability of edge to reduce
latency down to milliseconds, which is useful for certain applications such as augmented reality,
mission critical IoT systems or drones.

However, there are other benefits too, demonstrated in the table below, such as reducing network
backhaul and the bandwidth costs associated with this. This may be counter-intuitive for network
operators, but the reality is that in many instances data is not traversing the telcos’ wide area networks
and in fact being processed on devices or at premises. An example of this is in cars, which produce
tens of gigabytes of data each day; most of it stays in the vehicle. Edge computing will allow
automotive manufacturers to access data remotely and actually open up new connectivity revenue
streams for operators.
Figure 8: Benefits of edge computing

Hosting processing capabilities closer to the end user/device/source of data reduces round
Latency
trip times (down to a few ms) compared to having to backhaul data to the core
Since data transfer does not have to depend on the wide area network, there is a reduced risk
Reliability
that an edge application will be affected if there is a network failure
Processing data at the edge reduces the volume of traffic travelling through the backhaul
Reduced backhaul
network, reducing costs and avoiding strain on core servers
Growing concerns around data privacy, and regulations such as GDPR, make the ability to
Data localisation
process and store data locally (e.g. on-premise) key to some use cases
In contrast to dedicated appliances, edge compute infrastructure can host a variety of

Flexibility
applications and can quickly scale up/down or prioritise traffic based on demand
Offloading compute from an end-device to the edge allows this device to be both physically
Light device
lighter and consume less power, enabling otherwise inefficient use cases
Distributed applications across neutral edge infrastructure have workloads that can follow
Mobility
mobile end-devices, giving them access to compute wherever they need it
Applications are hosted on shared edge infrastructure so workloads can be easily moved to
Resilience
alternative servers in case of hardware failure, maintaining service continuity
To engage customers, telcos need to understand why and how each of their business drivers differ.
For instance, in certain customer segments - manufacturing, government, education, etc. – keeping
data on-premises and ensuring it is totally secure is paramount. In others, this is less of the case but
the reason for leveraging edge computing may be more weighted on the flexibility of the platforms it
brings. The retail and leisure sectors are strong examples of this, as they need to innovate to remain
competitive and continuously enhance the customer experience.
2. Select use cases

Use cases are like currency in telco edge computing and, although the telecoms industry dreams of
it, the killer use case does not exist. As with cloud, edge computing will support many different use
cases. We have already started to see some telecoms operators invest in edge use cases, for
example SK Telecom exploring MEC for augmented reality and AT&T looking into on-premises edge
for retail use cases.

For telcos, the number of potential use cases can be overwhelming, but it is important to focus on a
small number of use cases to start with, to begin to engage customers within specific industries and
build a solid minimum viable proposition that is relevant to at least some customers, and then begin
to develop new services. A telco could consider the following factors when prioritising a set of initial
use cases: its right to play, the size of the industry in its market, the value of edge in the use case and
the use case’s level of maturity. Below are examples of edge use cases evaluated by the incremental
value edge computing brings and the use case’s maturity (from a technology and adoption
perspective).
Figure 9: Example use cases evaluated
Use cases Value of edge Use case maturity

AR/VR field assembly and safety
Connected car driver assistance
Edge ADN/CDN
Private cellular networks
Security – video analytics
3. Define the business model

In 2017, we outlined the following business models in an HPE-sponsored paper ‘Edge Computing: Five
Viable Business Models’ and we have found that the business model landscape for telcos has not
changed dramatically. These business models do not define the business case for telcos deploy edge
infrastructure at their network edge sites; these are a given, since the sites are being used for
virtualised network infrastructure. Nevertheless, telcos need to be able to monetise the use cases that
will leverage the edge and be clear about where they seek to add value for providing edge-enabled
services.
Each business model has different implications on the addressable market and potential customer
base. Some telcos may choose to focus on dedicated edge hosting, primarily leveraging their network
edge sites for co-location services to cloud providers, brokers and large enterprises or providing bare
metal servers to run their (customers’) applications. Other telcos may prioritise taking edge-enabled
applications to end-customers (B2B(2X) solutions or end-to-end consumer applications), by either

building their own or partnering with ISVs to do so. Telcos are also evaluating providing services in a
cloud-like business model via consumption-based models, be they IaaS, NaaS or PaaS1.
Figure 10: Edge computing business models
Telco as enabler Telco as end-to-end provider
End-to-end
Dedicated Edge IaaS / Systems B2B2x
consumer
edge hosting PaaS / NaaS integration solutions
application
Platform business Project-based Solutions and

Co-location (space / Applications for the
models to provide solutions to build services for the B2B
power / cooling) consumer market
edge infrastructure, edge infrastructure market (end
and/or bare metal that leverage edge,
networking, tools as and applications for customers, ISVs or
infrastructure e.g. cloud gaming
a service B2B customers solution providers)
4. Co-create with the ecosystem

Telcos creating any edge-enabled services will need to build relationships with key partners. It’s almost
a given that partners will be critical to provide telcos with the components to build the edge site,
hardware and software platform and manage the edge platform. However, most value from edge
computing will come from the applications on top. Each use case has a different ecosystem of
partners that come together to provide an edge-enabled solution.
A key factor for telcos monetising edge computing will be to develop processes to scale these
ecosystems efficiently and be able to manage and onboard partners quickly. One way to do this may
be to build strong relationships with key players in target industries and work with them to co-develop
an ecosystem and tap into one another’s existing relationships. Telcos will need to carefully think
through the value proposition for each player in the ecosystem; why a gaming developer wants to
leverage edge from a telco in their game is different from the value derived by the end-gamer playing
the edge computing-enhanced game.
1 IaaS: Infrastructure-as-a-Service: compute, storage and networking in an as a service model

NaaS: Networks-as-Service: providing network services in a configurable, customisable way using virtual network functions
PaaS: providing additional consumption-based services to support IaaS, such as developer tools, application lifecycle management or
industry-specific platforms to help developers more easily build applications in certain domains

It is important to emphasise how early the edge computing market is and therefore how important it
is for telcos to seed the market by working with customers and partners to co-create new services.
This means fundamentally changing how telcos go to market. Rather than having a pre-packaged
product, telcos will need to follow a process of taking a minimum viable product to customers, working
with them to define the final solution, selecting the right use cases and application partners, and
defining the commercial model, consumption models, etc. Without this mode of working, it is more
likely a telco will place its bet on the wrong solution, invest heavily and fail to see it adopted in the
market.
5. Scale the business effectively

As telcos validate use cases and develop their business models, they will need to scale the edge
computing business. Part of this is about finding commonalities across use cases, be it types of
customers or the nature of the application. At a high level, some edge use cases have similar
workflows: collecting data, analysing patterns, identifying anomalies and triggering an automated
action. This is the case in using video analytics to identify suspicious behaviour in real-time, or when
analysing the condition of a machine for predictive maintenance. Telcos, therefore, can develop
certain skills and capabilities to offer services that support groups of use cases.
The role of orchestration

Scaling the edge business also means managing the complexities of edge computing infrastructure
and applications. There will be thousands of edge nodes that will take different guises – some
distributed compute will live in large telco data centres, some on small devices at the customer site.
Device and application lifecycle management (ALM) are therefore much more difficult, compared to
the cloud, because the infrastructure is so variable. In ALM, determining how well the application is
operating requires analysing multiple factors that are unique to how (and where) the application’s
workloads are deployed.
Another source of complexity is that the use cases’ workloads vary in their compute profiles. For
instance, cloud gaming workloads require high performance graphic processing that peaks during the
evenings, whereas in healthcare, the use cases’ computing needs are more regularly spread
throughout the day (e.g. monitoring a patient 24/7).
One of the main challenges in scaling edge compute will be in coordinating the right networking,
compute and storage capabilities to serve the needs of the customer or specific use case. Dynamic
orchestration of workloads and network functions will be key to ensuring the benefits of edge
(particularly latency) are met consistently and provide the customer with an assured quality of
service. This is especially difficult in mobile edge computing when applications will need to be
accessible as the end-user (person or machine) moves and workloads must be orchestrated across
edge sites.

Use case deep-dive

In order to kickstart the process that telco will need to go through, evaluating use cases and defining
business models, we have deep-dived on three use cases in this report: low-latency cloud gaming,
security video analytics and advanced predictive maintenance. Each use case will likely leverage a
different type of edge and create value for a defined customer segment.
Figure 11: Each edge use case leverages different benefits and may use a different
type of edge location
Advanced predictive
Low-latency cloud gaming maintenance Security – video analytics
Latency
Edge computing benefits
Reliability
Reduced backhaul
Data localisation
Flexibility
Light device
Mobility
Resilience
On-device
location
Edge
On-premises
Network
Telcos may also explore the potential of addressing vertical sector domains with a variety of use
cases, for example smart healthcare. This allows telcos to focus on building expertise in a single
sector and streamline investments into application platforms and services for that industry too.
Although this is broadly different from how telcos serve customers today, through horizontal services,
it will be a key requirement to be able to build relevant, edge-enabled solutions.

Cloud gaming involves hosting and rendering games in remote data centres rather than on devices
(PCs or consoles) and streaming the content to the client, i.e. the end-device. Doing this changes the
business model for gamers – rather than investing in owning the games and the hardware to run the
game, these could both be accessed via a subscription model. In a “Netflix for games” model, the
gamer does not have to download every single game they wish to play or spend thousands of dollars
on a high-end PC or console but can play across devices and minimise costs.
To be able to recreate the PC/console experience from the cloud, there needs to be high power
compute and low latency streams. Gamers are not tolerant of lengthy lag or ping times (latency in the
gamer world) and will pay a premium for high quality network services. A ping time of 20 milliseconds
roundtrip is good for a high-end game, 20-100 milliseconds is about tolerable, but anything above 100-
150 milliseconds is not good enough. Streaming high quality games from the cloud is therefore

impossible. This is where edge computing clearly plays a role, by reducing the roundtrip time between
the server running the game and the end-user.
In markets where fibre connectivity is not widely available, edge computing will allow cloud gaming
service providers to introduce high-end gaming where it was not previously possible by ensuring low
latency using wireless alternatives. In practice, this will open up the addressable gaming market in
these countries, particularly if the cost of owning a PC/console is too high for the general population.
Edge, combined with cloud commercial models, will allow these gamers to access high-end games
using simple end-devices.
Telecoms operators will need to think about their business models and who their customers may be
in edge-enabled low-latency cloud gaming, as each customer segment would be looking for a different
value proposition from the telco:
• Gamers (B2C – business to consumer model)
− Cloud gaming platforms: providing cloud gaming services to gamers either by developing the
telco’s own cloud gaming platform hosted at the edge, or distributing a third party’s
− Premium connectivity services: bundled with the gaming service or kept separately, telcos
can offer low latency network services (fixed or mobile), in the future using network slicing to
guarantee service quality
− Virtual PC: some gamers may prefer to buy games separately, but still want a cloud gaming
service that offers the ability to rent a PC in the (edge) cloud to avoid paying upfront for an
expensive device
• Game developers (B2B2C – business to business to consumer model)
− Distribution: (smaller) game developers will value partnering with telcos to access a
customer base and be part of a wider gaming platform, plus leverage telcos’ other assets
such as direct carrier billing
• Publishers, aggregators and cloud gaming platform providers (B2B2C)
− Distribution: similar to developers, third party gaming platforms may want to extend reach by
marketing and selling their services through telcos
− Edge cloud hosting: to optimise how the games on the platforms run, these types of
companies may leverage telcos’ edge clouds to improve performance
− Micro-services: these are smaller services that optimise gameplay, for example
matchmaking, which involves using location data to match gamers to play together to
ensure low latency
• Game engines (B2B2B – business to business to busines model)

− Performance APIs: games engines are the platforms developers use to build their games,
therefore telcos could leverage edge computing to provide APIs that use statistics on the
network/edge cloud to automatically adapt how the game runs to optimise performance
The real opportunity for telecoms operators comes in developing multi-sided business models, but
this requires building an ecosystem. Given cloud gaming is still a nascent opportunity, it is not clear
who the big winners are. Companies that telcos will likely partner with include cloud gaming platform
providers. Today, they are often focused on working with telcos to distribute their games rather than
develop deeper relationships that include leveraging edge computing. Telcos will need to drive a
change in the nature of these engagements to ensure they can add new sources of value.
Figure 12: Potential partners in cloud gaming
NVIDIA GeForce Now Google Stadia

NVIDIA will deploy their RTX Servers at the edge in telco networks Google Stadia will leverage Google Cloud’s edge nodes in order to
in order to enable latency-sensitive 5G and AR/VR cloud gaming offer superior latency
Shadow Tech Microsoft xCloud

Shadow Tech offer a virtual PC service, which utilises edge Microsoft’s cloud gaming service will leverage Microsoft Azure’s
compute specifically designed for low-latency data centres across the globe
Blacknut EdgeGap
Game streaming platform with a large catalogue of mid-tier Canada-based start-up looking to help games studios leverage
games in its portfolio edge infrastructure to solve latency problems and improve
matchmaking
Swarmio Media Hatch
eSports platform-as-a-service who want to engage the telco Mobile cloud gaming service, already partnering with telcos such
community and leverage edge computing for eSports as Vodafone to provide a cloud gaming platform
Video analytics – security

Video analytics is a huge area that can benefit from edge computing; the global market size for video
analytics is predicted to reach USD 20 billion by 2027. Applications are data-hungry and use significant
bandwidth and there is an increasing need to use video to conduct real-time analytics to trigger an
immediate response.
This creates a challenge for using the cloud: it’s too expensive to transport the data, given that an
individual high quality camera generates at least 5 megabytes of data per second, which scales up to
1.08 terabytes per day if using 25 cameras at a premises. In addition to the network costs,
continuously having to store the accumulated data creates huge cloud storage costs. Finally, it slows
down the business process, primarily due to latency, but also due to analysing all the data at a
centralised location.

Manufacturers such as Bosch and Avigilon are developing edge compute capabilities on the camera
itself. Although there is an advantage for the customer in eliminating network costs, it drives the cost
of the end-device up and requires customers to have to replace their existing camera real estate. Plus,
end-devices will have limited compute available. Edge computing (on-premises or on the network)
allows enterprises to leverage video analytics applications in a scalable and more efficient way.
In this paper we are focusing on security use cases, which include using real-time analytics to detect
suspicious behaviour, accidents and identify objects. Part of the need for edge is to access these
applications that allow governments and enterprises to ensure secure environments more effectively
and act quickly when there is a risk. However, the other benefit edge computing brings is to filter out
data that is not required for further analysis. For example, reducing the need to store footage from a
security camera at a car park during the night when no one (or thing) has appeared in sight.
Outside of security and surveillance, video analytics solution providers are using similar capabilities
and edge infrastructure platforms to enable other applications. One example here would be for use
cases that require the number of individuals in a given location to be counted. This is a huge
opportunity at the moment in the COVID-19 environment, as businesses and government are
increasingly using technology to ensure people comply with social distancing and other policies. For
example, Uber is exploring the use of video to monitor if drivers and passengers are wearing face
masks. Other examples of video analytics being used in different domains include: footfall analysis (in
retail) to understand behaviour patterns in shoppers; in building management, it can be used to
manage heating and lighting controls automatically to save energy; in mining, it can be used to detect
individuals in hazardous environments and trigger a warning.
Video analytics for security is a compelling opportunity area for telcos; it leverages telco capabilities
in edge and secure networking, and is applicable across multiple industry sectors. However, telcos
who pursue this is an opportunity area will likely need to define target customer segments with specific
value propositions to avoid solutions and services that are too generic and do not differentiate from
those that exist on the market already. As ever, partnerships will be key and telcos may want to
consider a range of partnerships, including (local) start-ups who are creating innovative software
solutions, as well as established security systems integrators and installers. Finally, telcos should try
to provide services that work across various connectivity options, not just cellular, given that
customers will own cameras that are already connected by different technologies.

Predictive maintenance refers to the process of monitoring the key parameters of an asset and
applying machine learning or artificial intelligence to determine when the machine needs maintenance
to avoid it breaking down. This has significant benefits in making maintenance and operations more
efficient, reducing equipment downtime and improving productivity.
It has been a concept in industrial segments such as manufacturing for many years but has faced
challenges in customer adoption. Part of the issue has been that enterprises still face difficulties in
extracting data from machines, systems and processes and using cloud based IoT applications to
analyse the data.

Edge computing can reignite predictive maintenance (hence advanced predictive maintenance) by
allowing data to be processed at the customer site, as close to its source as possible. It is too costly
to install thousands of sensors across hundreds of machines to monitor every single parameter
(humidity, oil leaks, temperature, etc.) if the data is analysed in the cloud. Plus, many enterprise
customers in industries such as mining, manufacturing and oil and gas are reluctant to use the (public)
cloud for processes close to operational technology (OT) due to sensitivity around the data. Edge
computing also uses standard hardware and software to make it easier to innovate, adapt applications
and integrate into existing enterprise systems.
A key trend happening in these industries is the convergence of IT and OT, part of this includes using
standard IT software, hardware and processes within core industrial operations. This is part of the
general move towards Industry 4.0; IoT, edge and cloud are all tools to enable this digital
transformation. Some industrial segments are particularly interested in leveraging 5G (and LTE) to
introduce more automation in the plant using IoT sensors and robotics.
Telcos will need to understand the complexities of providing an edge-enabled predictive maintenance
solution. Some of the potential ecosystem partners are shown below. Operators will need to think
about their role and how they create effective partnership models to take solutions to customers with
these companies. On the one hand, the telecoms operator may want to front the customer solution,
providing some elements such as the connectivity, edge cloud platform, IoT and analytics platforms,
but using partners to provide the applications, hardware and systems integration components. In
other cases, the telco may prefer to provide services for others to take solutions to market, for example
selling networking and compute to a systems integrator as part of their end-to-end solution.
Figure 13: The ecosystem required for advanced predictive maintenance
Schneider
Bosch Electric
PTC
OEM / Machine
machines IoT platform controller
Litmus (PLC)
Edge
analytics Telecoms
platform operator
Salesforce
CRM
Edge server Orchestration
Rockwell
Automation
Systems
integrator

The ecosystem development process will include defining the customer proposition and
understanding how these solutions help to solve business problems. Three primary customer
segments for advanced predictive maintenance are heavy industries (manufacturing, oil/gas, utilities),
transport and building management. For instance, each one has a different business case for why
downtime is so critical, therefore why predictive maintenance would benefit enterprises:
• Manufacturing: Unplanned downtime costs manufacturers an estimated $50 billion annually,

with 42% of this caused by equipment failure
• (Air) transport: The cost of aircraft downtime estimated at $10,000 per hour
• (Rail) transport: Maintenance accounts for around 50% of overall asset cost
• Building management: Elevators worldwide are reportedly ‘unavailable’ due to maintenance for
190 million hours yearly
Smart healthcare
The last deep-dive centres on a vertical domain to explore how telcos can serve a single segment with
multiple edge use cases. In general, telecoms operators are starting to verticalise and invest in specific
industries. This is because, in order to provide services beyond connectivity, there is a need to have
specialised skills and capabilities that are relevant for developing platforms, applications and solutions
for that industry’s ecosystem. Outside of certain areas such as cloud, networking and some IT
systems, much of these are specific to a vertical.
We have already seen telcos start to prioritise healthcare as a vertical, for example Telus Health,
Telstra Health and KPN Health. It lends itself to operators, given that there is an increasing need for
(5G) connectivity and cloud as the industry transforms, adopts IoT, moves away from manual
processes, etc. Plus, healthcare provision and the value chains are highly local by nature, therefore not
always easy for global companies to address.
With edge computing, there is an opportunity for telcos already investing in healthcare to expand their
play and introduce new applications, solutions and services to healthcare providers. For others, who
may only be serving healthcare enterprises with networking services, edge computing could be a way
to deepen relationships and start to build a vertical play.
Healthcare is a large domain spanning across hospitals, care homes, pharmaceuticals and broader
wellbeing. Across these, rising demand for healthcare services driven by ageing populations and new
diseases (notably COVID-19 recently), plus limited resources and growing costs related to improving
healthcare outcomes, mean that technology can play a role in increasing efficiencies while improving
outcomes:
• Offering higher quality care for patients with improved ability to respond to critical situations
• Increasing the productivity of existing systems and resources

• Reducing the burden on healthcare services by mitigating illnesses through wellbeing services
• Making better use of patient data and improving data mobility, while maintaining its security and
complying with data privacy regulation
The main benefits edge computing brings is allowing healthcare providers to access new (IoT)
applications that were challenging to adopt in the past. These are being used to monitor the wellbeing
of people in their daily lives to reduce the chances of them needing to go to hospital, making in-patient
care more effective and efficient and supporting remote care.
However, most applications are built on the public cloud, which is problematic because is it not easy
to manage data and ensure it is stored close to the individual. Edge computing makes it easier to keep
data secure. It also unlocks the ability to use data analytics to make real-time decisions or trigger
alerts, for example when conducting real-time scans or scanning for incidents that may be mission
critical to the patient’s health if not acted upon immediately. These require low latency processing that
edge computing enables. Some of the specific use cases being explored are outlined below; in certain
situations (e.g. if the use cases requires data to be processed on-site) an on-premises edge at the
hospital is the most likely option, whereas for remote care a network edge is more appropriate.
Given this, telcos can be in a strong position to orchestrate the edge. This includes determining which
edges are required, ensuring the right workloads and applications are being run from the appropriate
edge, determining how the network can be optimised to deliver the demands of the use case and
incorporating partner services to enhance performance.
Figure 14: Smart healthcare use cases

Relevant edge factors
Use case Latency Reliability Reduced Data Flexibility Light Mobility Resilience Why edge?
backhaul localisati device
on
Data sovereignty maintains security of patient

Connected data, whilst low latency enables real-time
ambulance analysis of patient vitals to increase efficiency
of responses in-vehicle and at the hospital
Low latency to quickly detect and act on
emergencies, but also providing services in a
Remote care
scalable way without needing
heavy/expensive devices and equipment
Depends on reliable processing of vast
Video analytics amounts of data (from video feed): edge
enables quick analysis for access control
Low latency can be crucial to respond to
In-patient
anomalous data that could be critical, whilst
monitoring &
processing data on-premise maintains
diagnosis
security of sensitive data
Data stays local (rather than being processed
Data mobility in a central cloud) to ensure compliance with
hospital and government regulations
Latency is important to ensure real-time
interactions with patients; will gather sensitive
Smart robot
patient information that does not need to
travel to a central cloud to be processed

There are examples of telcos and solution providers already experimenting and enabling such use
cases:
• Vodafone conducting proofs-of-concept with Nokia in Italy for connected ambulance
• SK Telecom using AI and video analytics for multiple use cases, including contact-free entry
using facial recognition
• HPE and Cerner developed a solution to detect sepsis early when patients are in hospitals
For telecoms operators considering their strategy in edge-enabled healthcare, they will need to
understand which edge solution is most appropriate and who are the partners and customers to
engage. Hospitals have existing relationships with vertical specialist OEMs (e.g. Cerner, Siemens,
Philips, General Electric, etc.), therefore any go-to-market plan will need to outline the dynamics of the
telcos’ relationship with these established players. In addition, applications are being re-architected
from on-device processing to leveraging the edge and, again, these device manufacturers are key
stakeholders.

Next steps
This report sets out five key principles telcos should consider when defining a strategy to monetise
edge computing, as well as potential use cases. However, edge computing is a nascent market and
any party wishing to take a stake must realise that it needs to play a role in developing the market to
solidify a position in the ecosystem going forward. In practice, this means taking an iterative approach
and working directly with partners and customers sooner rather than later.
Taking the three use cases and the smart healthcare vertical domain as examples, this is how a telco
may want to think about applying the principles:
Figure 15: Five principles applied to use cases
1. Learn edge 2. Select use 3. Define 4. Co-create 5. Scale the

Use case
benefits by cases business with the business
domain
industry models ecosystem effectively
Reduce latency/ping/ Related use cases Applications: gamers Content is king; Consider the
lag could include IaaS/PaaS: gaming studios, aggregators orchestration
Reduce end-device matchmaking platforms and developers with capabilities required
Low-latency TCO players, AR/VR Dedicated edge the best games are to provide edge-
cloud gaming Enable new types of gaming hosting: cloud critical enabled cloud
games and models providers Develop potential gaming across
platforms / APIs with multiple devices and
games engines networks
Reduce backhaul and Related use cases Dedicated edge Work with ISVs that Select one industry
TCO for customer could include hosting or IaaS/PaaS: are specific to each segment to work with
Enable new business condition-based Industrial OEMs/SIs industry segment and to start with and build
Advanced models for monitoring and ISVs SIs to integrate into MVP with sensor
predictive maintenance Solutions: industrial CRM / ERP systems, manufacturers and
maintenance Gain customer end-customers leverage existing IoT ISVs, leverage
insights relationships where existing IoT
possible platforms, define role
of network edge
Reduce backhaul Sub-use cases Dedicated edge Start with single Start with single
Allow customer to include real-time hosting or IaaS/PaaS: security use case and security use case and
access innovative object/event Security SIs and ISVs develop MVP from develop MVP from
Security – applications without identification, health Solutions: enterprise there, then work with there, then work with
video changing and safety end-customers customers to customers to
analytics infrastructure understand how this understand how this
can be extended to can be extended to
other video analytics other video analytics
applications applications
Access to cloud Sub-use cases IaaS/PaaS: IoT / Need to work with Ensure access to
services without include: remote application providers established remote patient
using public cloud patient monitoring, in- Dedicated edge healthcare specialist monitoring
Enable remote patient patient monitoring, hosting: SIs or private SIs and OEMs that applications by
Smart electronic healthcare healthcare providers orchestrating
monitoring cloud providers
healthcare records trust and leverage workloads across
Maintain data privacy Applications /
solutions: consumer local (public sector) multiple edge sites
or healthcare relationships within for ubiquitous
providers market population coverage
Even if a telco does not choose to pursue the use cases above, there are immediate next steps
telecoms operators should consider regardless of this:

1. Learn the edge benefits for each industry: Build initial propositions to engage customers
and partners to understand how they benefit from edge computing and work together to
create prototypes to validate the business case.
2. Select use cases: Given that edge computing is still evolving, it does not make sense as a
telco to try and predict the future. Telcos should quickly seek to prioritise a small number of
industries or use cases, which could be based on where there is most value from edge and
where the telco has a right to play – e.g. a large industry in its market. These use cases will
serve as a starting point to conduct proofs-of-concept with customers and application
developers, while taking an iterative approach to adapting the priority use cases.
3. Define the business model: Identify potential areas for IP/value creation and base the
business model on that; do not try and control the whole stack – learn from the mistakes of
the cloud! This exercise will require a better understanding the value chain in a specific
market to identify where the telco can add value and then apply it to other verticals and use
cases.
4. Co-create with the ecosystem: Work with customers sooner rather than later, taking an
initial minimum viable product to market to define the commercial models. This process may
require training sales/pre-sales teams to be more consultative in their approach. This is
fundamentally different to the telco sales model today. To build the partner and customer
ecosystem quickly, telcos must make it easy for parties to access edge computing and new
network services to accelerate the innovation process and go-to-market. This could include
creating common, easy-to-use interfaces for partners, using APIs, making sure on-boarding
is seamless, being able to orchestrate their capabilities readily into services and labs, demos
and self-serve platforms or marketplaces to promote services and applications.
5. Scale the business effectively: Begin to build a solid foundation for scaling use cases,
services and infrastructure. In product development, find commonalities and capabilities
across use cases that can be developed into services. On the infrastructure side,
orchestration capabilities will be needed to optimise the distributed edge node architecture
and best serve the various workloads that they will support. Without dynamic orchestration,
there is a risk that edge nodes may be underutilised and economies of scale will be
impossible to achieve.

Research Consulting Events
www.stlpartners.com

Business White Paper - Telco Edge Computing

Uploaded by

Copyright:

Available Formats

Business White Paper - Telco Edge Computing

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Business White Paper - Telco Edge Computing

Uploaded by

Copyright:

Available Formats

Executive Briefing Service

Kindly supported by:

TELCO EDGE COMPUTING:

Dalia Adib, Practice Lead – Edge Computing | dalia.adib@stlpartners.com | October 2020

Figure 1: 5 principles for commercialising edge computing

Source: STL Partners

© STL Partners EXECUTIVE BRIEFING 2

Use cases Value of edge Use case maturity

Connected car driver assistance

Low-latency cloud gaming

Private cellular networks

Advanced predictive maintenance

Security – video analytics

Source: STL Partners

• Low-latency cloud gaming

• Security video analytics

• Advanced predictive maintenance

Next steps for telcos: putting a monetisation strategy into action

© STL Partners EXECUTIVE BRIEFING 3

Figure 3: Five principles applied to use cases

Source: STL Partners

© STL Partners EXECUTIVE BRIEFING 4

Introduction to telco edge computing ................................................................................................................ 7

Key principles for telcos seeking to monetise edge computing ................................................................... 9

Use case deep-dive .............................................................................................................................................. 14

© STL Partners EXECUTIVE BRIEFING 5

Figure 3: Five principles applied to use cases .................................................................................................. 4

Figure 4: Telco edge computing locations ........................................................................................................ 7

Figure 5: Virtualised RAN architecture uses edge computing ....................................................................... 8

Figure 6: The edge computing value chain ....................................................................................................... 8

Figure 7: 5 principles for commercialising edge computing .......................................................................... 9

Figure 8: Benefits of edge computing .............................................................................................................. 10

Figure 9: Example use cases evaluated ........................................................................................................... 11

Figure 10: Edge computing business models ................................................................................................ 12

Figure 12: Potential partners in cloud gaming ................................................................................................ 16

Figure 14: Smart healthcare use cases ........................................................................................................... 20

Figure 15: Five principles applied to use cases .............................................................................................. 22

© STL Partners EXECUTIVE BRIEFING 6

Introduction to telco edge computing

Figure 4: Telco edge computing locations

Enterprise Radio Access Network Metro Core Network Internet

Telco edge computing

“Far edge” Network edge

Source: STL Partners

Initiative Relevance to edge computing

© STL Partners EXECUTIVE BRIEFING 7

Figure 5: Virtualised RAN architecture uses edge computing

Source: STL Partners

Figure 6: The edge computing value chain

Edge cloud Systems

Source: STL Partners

© STL Partners EXECUTIVE BRIEFING 8

Key principles for telcos seeking to monetise

Figure 7: 5 principles for commercialising edge computing

Source: STL Partners

1. Learn the edge benefits for each industry

© STL Partners EXECUTIVE BRIEFING 9

Figure 8: Benefits of edge computing

In contrast to dedicated appliances, edge compute infrastructure can host a variety of

Source: STL Partners