3 Choices on how to do Telecom Edge

Building edge is not like building cloud. Edge is physical and is location-specific. It has a finite capacity and must be planned to maximize utilization and minimize footprint, power, and cost. Building edge is more like building a Radio Access Network (RAN) in the mobile industry. The existing RANs that we all take for granted today, and which enable our smartphones to work at all, are already the largest global edge computing network in existence and will remain to be so in the future. They are just dedicated to one application - wide-area mobile connectivity for devices.

In parallel to 5G investments and architectures, there is an opportunity for mobile operators to build a multi-purpose edge computing platform, not only for their applications but for other people's applications. This makes common sense since it can inherit the existing locations, operations, and presence the mobile operators already have. However, this does not appear to be happening in a focused predictable manner.

"So there are multiple options for operators to consider: But they need to get past the FUD stage and quick, otherwise they'll find others eating their lunch (again) and their 5G strategies crawling. For those not already committed to an edge computing plan, it's time to take some risks."

- Ray Le Maistre, LightReading,

"FUD Still Weighs on Telco Edge Computing".

With regards to telecom operators, this fear, uncertainty, and doubt is passed on to the market and thus colors the potential demand away from telecom to other solutions. Telecom operators need to have a clear plan for 5G and within that a telecom edge.

There are lots of choices on how to build a telecom edge, let's simplify it to three and add a simple decision-making process on when to use one versus another. This then allows a 5G edge computing strategy, that is unique to each telecom operator, depending on the very different aspirations, capabilities, and regional presence.

Reference Model for this Discussion

Three models are discussed, from the perspective of the telecom operator.

1. Shared Services
2. Shared Infrastructure
3. Shared Nothing

With 1. Shared Services, the telecom operator offers its own platform to their industries of choice. Since each operator only represents a portion of even the markets they operate in, this option requires a level of federation across all operators, first in a country and then most likely globally. This is the same type of federation we are familiar with in the telecom business of today, where the voice experience is shared across customers of all carriers. The telecom operator has direct ownership of the customer. MobiledgeX Edge-Cloud is a good example of an independent platform that is working on behalf of telecom operators to both enable direct customers and also the option 2 for monetizing shared infrastructure with customers the operators do not target and will not target.

With 2. Shared Infrastructure there is a proprietary platform that one or more applications code against. The platforms run on a common infrastructure that the operators sell to the platform providers. The different platform providers may be industry-specific or in the case of Google Anthos, be public cloud extensions that run over any infrastructure. Direct customer ownership is with the proprietary platform. The telecom operators accelerate the platform providers' ability to distribute in their market, as growth occurs.

With 3. Shared Nothing the proprietary platform also vertically integrates with a proprietary control plane and potentially a proprietary infrastructure, that is dedicated to the one platform. AWS Outposts and Azure Edge Zones are good examples of this model and really serve as a multi-tenant dedicated appliance model. The telecom operators provide connectivity to the edge zones, the platform providers handle all other aspects.

The red line indicates what added value is being sold. 1. Services being the highest potential value, 2. Shared (location specific) Infrastructure being a new value and 3. Shared Nothing indicating connectivity, which is the same business model as of today and no increased top-line value. As will be discussed later, all models are good decisions to take in the right context.

Building the Correct Strategy

Start with the customer in mind...

Software developers are the customers of any edge. The software developers that need edge are those building a real-time world understanding that affects the experience being delivered. It is really important to understand the actual customer, the developer. The worst possible edge to build is one that attracts no customers. And developers, just like customers anywhere, come in different shapes and sizes.

Question 1 - Which developer customers do you wish to serve?

Which model is best to use when?

The market is not a uniform distribution of common needs or maturity and a telecom business model is balanced across, new revenue generation, customer experience, and delivery efficiency. All 3 strategies have a different role to play in different parts of the market segmentation.

Consumer/Existing Application - Most Interesting

These are the applications that have built the app economy. They are all born and live in the public cloud. Some have moved to their own infrastructure when scale and predictability have enabled the economic justification for their own operations. Many of the largest applications are owned by the public cloud providers, for example YouTube. Collectively, they represent the bulk of the traffic traversing any telecom operator's network. They are quite often very happy being married to one public/own provider since it has proven to be the simplest and fastest path to market and there are no business reasons to not do so.

These applications will increasingly create new real-time experiences that will increasingly require edge computing to work well. They will not generate net new money through doing this (many are free, freemium models already) but they will drive engagement (their business model), which allows them to monetize indirectly.

From the telecom perspective, there should not be expectations of generating interesting net new revenue from this space in the short term but the experience from these applications will continue to define the customer experience of operators customers as it does today. Those operators that enable the best end-user performance while keeping delivery costs to a minimum, will be best positioned to claim larger market share as 5G is being deployed and not all operators are able to deliver such comparable experiences.

Recommendation

The best deployment model for this is "2. Shared Infrastructure", where possible (example Google), and "3. Shared Nothing" where not (example AWS, Microsoft).

Enterprise/New Application - Most Interesting

This is the greenfield market opportunity that everybody is excited about, the enablement of the real-time enterprise. This is the processing of large amounts of time-sensitive data to create real-time hyper-personalized experiences and highly efficient automated supply chains. It is the application of video analytics, industrial IoT, and robotics, automation. The current crisis has accelerated all enterprise digital transformation plans and has moved them from something that can happen in the future to something that must happen now.

Enterprises, like operators, do not want (and in some cases are not allowed) be locked into single suppliers, so the software companies (Independent Software Vendors) focusing on this digital transformation must do so in a way where they maintain cloud provider independence. Enterprises also have strong governance, security, and data residency requirements, all buyer criteria that play well into the telecom industry promise.

Looking through the eyes of the Independent Software Vendors (ISVs)

These companies are not focused on which technology to use, they are focused on generating business outcomes for their enterprise customers. If they can generate more business outcomes with better results, then they succeed, their customers succeed and their suppliers succeed. But the suppliers are last, not first in terms of the hierarchy of importance.

• What business outcomes does the customer need?
• What supplier requirements do they enforce on the ISVs?
• What performance guarantees and assurance is required?
• What solution life cycle management guarantees are required
• What software architectures best enable such solutions?

For more thoughts see here - "Edge Computing - How to Find Demand?"

Recommendation

The best deployment model for this is "1. Shared Services", where possible, and "2. Shared Infrastructure". The worst model here is "3. Shared Nothing" since the business model from the perspective of the telecom operator remains to be connectivity only and does not achieve increased relevance and/or revenue.

Consumer/New Application - Least Interesting

We will not dwell on these quadrants too much here. Consumer applications are the domain of the public clouds where rapid prototyping with low-risk investment (a credit card) is enabled. Because of the clear attachment to the public cloud developer environment, it is hard for telecom operators to build new business. If the strategy for "Consumer/Existing Application" has been implemented, the public cloud edge presence will also satisfy these developers and create maximum efficiency.

Enterprise/Old Application - Least Interesting

The majority of these applications are the ones that should move to a centralized cloud SaaS model anyway since unless they are OT rather than IT, there is no need for their edge presence. The companies best positioned to help companies with these are the consulting firms advising on total digital transformation strategies.

What Would I do as a Telecom Operator?

Believing there is one strategy to rule them all does not accommodate the simple fact that there are many developer customers serving many different markets that all have different requirements, status, and maturity. And also to believe that the advantages, disadvantages, and existing strategies within all operators is equally naive.

I believe the best strategy is to understand which market segments each operator wants to address directly, for the unique strategic advantages and reasons known to that operator such as government alignment, existing customers and partnerships, and existing business goals and revenue targets for 5G. This strategy should give the execution roadmap on how to validate/succeed with the already existing business goals of the organization, as well as setting its own.

Once the known direct coverage, across both enterprise and consumer, is known, an execution matrix across partners can be put in place, with appropriate priorities. The same can be repeated for the indirect market and similar for the "everything else" market.

Execute, trial, validate directly with customers to ensure your understanding of market is in line with their real business needs and plans. Doing nothing is not a strategy, doing everything is not either.

Other Considerations

Telecom operators are not just another company, they are a highly regulated, mission-critical service provider for all aspects of society, and telecom networks represent a strategic investment and area of interest for all governments. Other areas that must also be considered in any strategy include:

Sustainability. Global greenhouse emissions from the Internet are already estimated to be about the same as the airline industry before the pandemic and were already predicted to double by 2025. From a long term perspective, building a shared common infrastructure at the edge would appear to be more sustainable than building individual edges for individual providers. With cloud this has been possible since it is centralized, with the edge, where there will be 1000's of sites, it will be less so.

Supplier Diversity. The vast majority of telecom operators are required to have multiple suppliers for the same goods, thus building resiliency. Proprietary solutions and vendor lock-in are the antithesis of this. By building open platforms this allows a diversity of supplier base and de-risks any chosen strategy.

Conclusion

While the telecom edge landscape looks complicated initially, it does not have to be. The balance of the right strategy to the right market segment lays the foundation for a more targeted analysis of the options under the strategic light of each separate telecom operator.

Choose an "execute and learn" focused strategic execution, to help guide both 5G and edge computing investments, trialing the different partnership matrices required.

The future leaders of edge and 5G are doing this, join them, and design the next reality.

“In the face of the unknown, entrepreneurs act. Specifically they: 1) Figure out what they want, 2) take a small step toward making it reality, 3) think about what they learned from taking that step, 3) build that learning into their next step, In other words: Act. Learn. Build. Repeat.” – Forbes

About MobiledgeX: MobiledgeX has been working on solving these problems on behalf of the telecom industry for over 2.5 years. To learn more visit https://mobiledgex.com