The Future of Cloud Native Telcos

Originally published on CloudOps’ blog.

Ian Rae and Maddison Long discussed the future of the telecommunication industry. They focused on the impact of open source adoption, increased competition between cloud providers, and the ability of cloud native telcos to provide differentiated cloud and networking services in their target markets. Listen to the podcast below.


ML — What would you like a telco to take away from this conversation if you were to personify them as a person?

IR — That’s a good question, Maddison. I have a lot of personal history in telecommunications because my father was a telecommunications engineer. It was something I was brought up with. As I got into the world of the internet, I had many conversations, even arguments, with my father about how human beings will communicate in the future. In its early days, the conversation was mostly about circuit-switching versus packet-switching. It eventually became a conversation about cloud. My father was adamant that telecommunications had already invented the cloud and had been using the cloud as a term for many decades. I initially thought that was ridiculous, but it turned out he was right and was just referring to something a little different.

What I would most want a telco to take away from this conversation is how much I really believe the telco mission is unbelievably important and fundamental to how we will coexist and cooperate on this planet. While telcos all solve similar, fundamental problems, I’m amazed by the diversity that comes from them navigating the uniqueness of every region. That is an incredibly difficult and important mission.

As part of the so-called cloud native generation, I hope telcos will be able to successfully transition to this new world of software-defined services and everything-as-a-service, while taking advantage of their regional knowledge and existing infrastructure investments. This is important to get right, not just for telcos themselves but also for the people and businesses that depend on their services.

ML — Do you see a role for open source in that mission?

IR — Absolutely, especially when I think about this transition towards an as-a-service economy. We’re moving away from selling boxes towards a subscription economy. This is happening in every industry. This has a lot of consequences, not the least of which is depending on reliable, resilient, cost-effective, and secure cloud services.

We learned very early in the cloud computing industry that it’s basically impossible to build at scale cloud services without open source. At CloudOps we’ve gone a little further in that we believe you can maximally take advantage of cloud services. The cloud has won, which is the good news. The bad news is that all the clouds have won. You have to figure out how to take advantage of all the different cloud services to solve any business problem.

That introduces complexity and requires a layer that brings everything together. It’s incredibly important for our customers that the control plane being used by their services can rely on continued support in perpetuity. This is where we see the importance of the open source layer both for cloud computing providers and the ways we take advantage of cloud services. Most of these large open source projects are likely to survive vendors coming in and going over the market. This is something customers and service providers should take advantage of.

ML — While open source technologies have helped unshackle many telcos from traditional vendor lock-in, they can be complex to build, test, manage, and upgrade. Many have therefore moved away from vendor lock-in to being heavily dependent on a system integrator to execute the day-to-day operations of their open source-based platform. This is especially true in networking, where many open source technologies have large monolithic architectures. Have you seen these circumstances with some telcos you’ve worked with?

IR — As a technologist, I tend to see technical solutions when I look at problems. Open source software has a technical component, but also a business model and an operational model aspect. If we take a more holistic view of what is open, we will see that it is possible to implement open source in a way that is very closed both from a business and operational perspective.

At CloudOps, we would advocate that owning your destiny will be proportionate to whether the software you’re using has an open source licence that matches your use case, an open community that allows you to participate, and implementation that is in an open operational model. The intellectual property and know-how around implementing and operating open source must be accessible and its operation must be feasible for you as an organization. You will have lock-in in a different way if you instead depend on someone else to operate your open source software in a more black box manner.

This is important at a business model layer as well. You could quite literally build a solution based on open source that makes you depend in perpetuity on an integrator, for example, to deliver your services. The way we address this is by making sure that our operational support services have an open stance that can be operated by our customers themselves. We will literally go in and train our customers to do so by equipping them with all the playbooks, automation, and techniques that we use internally. Our mission is to enable them to do that. It is true that a lot of integrators and suppliers are not willing to go that far.

ML — Concerning these communities and their sustainability, there seems to be much criticism towards certain end users (who I won’t name) that leverage open source software to do their undifferentiated heavy lifting but then build and monetize differentiated value on top of these open source solutions. What responsibilities do telecoms and other end users have towards the sustainability of the open source community and the technologies they’re leveraging?

IR — That’s an important point. A key question I like to ask telcos is what are their differentiated versus undifferentiated services, which ties back to the general problem of knowing yourself as a business. What things, as a telco, are strategic capabilities that could be on the roadmap and have consequences for how you approach building the operational and technology layers of the business? This conversation or mapping exercise is a learning experience for everyone involved.

As we try to break down all the building blocks of the layers to be able to deliver service, it’s interesting to see how some building blocks are very clearly undifferentiated. The important piece becomes whether the service provider can be world class and operate the differentiated component. In doing so they can differentiate themselves as being a world class service provider. In this case, the burden of building that technology isn’t necessarily on the service provider.

In other cases the service provider has very differentiated services that make sense in their particular region and which they must invest in, build, and do R&D to control. These service providers must invest a lot of their developer resources in the exercise of mapping their differentiated versus undifferentiated building blocks, understanding how they ultimately get composed at a higher level into services that are outstanding and exceed customer expectations. While the customer doesn’t care whether the underlying building blocks are open or closed, they do care about having great service and value. Service providers can get there by really knowing themselves and knowing why which building block has been chosen. Because CloudOps is a big fan of open source, people sometimes assume that we think everything should be built out of open source.In reality, we see a need for a mixture of proprietary and open source. What really differentiates many of our customers is how cleverly they’ve selected which type of technology to use for one purpose and then combine with others into an overlying service.

ML — Without going down the rabbit hole of defining what it means to be a cloud native telco, what do you think telcos can do today to become more cloud native and differentiate themselves in their markets?

IR — It’s always fun to have new market buzzwords and then try to collectively figure out what those words mean. It’s a really interesting process that we use in our industry, but like any buzzword there’s fire behind all the smoke. For me cloud native telco evokes a few things.

First of all, cloud native certainly evokes the notion of meeting the fundamental characteristics of cloud services. What does it take for a telecommunications service to meet those characteristics? Is it self-service? Does it have transparent pricing/metering show back and charge back? Is it all software-defined and software automated? Does it have APIs that are first class citizens and that allow the products to actually be automated directly by customers, partners, and the telco themselves. Can these services be controlled through a complete and software-defined manner.

Another piece that cloud native evokes for me is the notion that the services being built can theoretically run on multiple underlying cloud architectures. That’s a very important and relevant part of the puzzle. A telco will typically cover a geographic region and a whole set of consumer and business interests. Every region has a different mixture of cloud service providers and capabilities for a variety of reasons.

Another key piece that cloud native telco evokes for me is the notion that one could design a system that fundamentally is flexible enough to run where the telecommunications service provider needs to run, whether on the edge, in their own infrastructure, on their own cloud, or in a public cloud, like Amazon’s AWS, Microsoft’s Azure, or Google’s Cloud Platform. That flexibility is an enormous opportunity for service providers.

ML — Many telcos and many analysts see the hyperscale cloud providers as fundamentally competitive. I’ve always seen this kind of growth in the hyperscale cloud providers as an opportunity for growth within the telco industry. How do you think telcos should interact with hyperscale cloud providers, and do you see them taking a more competitive or collaborative approach?

IR — I definitely hear anxiety and concern coming from telcos with regards to the hyperscalers. This is in part because a lot of their services are delivered over the top in a very telco-agnostic manner, which is a concern. It is also in part because some of the demands being driven over the top are having real world impacts on the underlying infrastructure. This can send a telco scrambling if they suddenly and unexpectedly have to change their network’s capacity model. That’s also a concern.

That said, I will reiterate my earlier point that while it’s good news that the cloud has won (certainly for companies like CloudOps), it’s bad news that all the clouds have won. The opportunity for telcos, as the regional gateways and delivery mechanism for these incredibly valuable cloud services, is to differentiate themselves through their ability to allow their customers to experience all of those cloud services better than any other competing telco.

It’s really problematic if you as a telco think that you have to pick Amazon, Google, or Microsoft as partners. That misses the point. Telcos must give their customers what they are asking for, which is efficient and reliable access to all the cloud services. As businesses are starting to rely on the cloud, connectivity is becoming much more important for them. This is great for telcos and could put them in a much more important position than they’ve ever been in, than a situation where there’s a large enterprise, everybody shows up, and there’s some local area networking and maybe some WAN networking that’s being depended on to deliver applications. Whereas in the cloud world your users are accessing applications that are all over the place and are moving all over the place. This is leading to some really interesting and challenging problems that telcos are in a great position to solve with solutions for the market, but only if they enable the customer to access to that whole menu of cloud services. I’m encouraging our customers to embrace all the clouds as they are ultimately driving the demand and the importance of the telecommunications sector.

ML — We’re seeing 5G emerge as a huge buzzword, and terms like 5G phones, 5G network, 5G use cases are rolling out. Do you think there are use cases that 5G will enable in the future beyond faster internet?

IR — Absolutely. We’re at a transition point in wireless where it would be arrogant for me to speak with authority as to what the most interesting use cases will be. I suspect that those will emerge once we start to see the real world capabilities. I actually think we’re at an incredible inflection point for innovation, and we’re going to see brand new things that we haven’t seen before as a result of 5G. I really believe it’s something new.

That said, the things I’m most excited about with 5G are the density, the reach, and, in particular, the enabling of some of these long distance, low power networking that were really not possible before. Also, there are very interesting backhaul architectures that will allow new use cases. This is where I think 5G will really help the edge explode. It may be strange for me to say this as a cloud guy because I’ve spent so much of my career evangelizing why it’s a great idea to centralize computing and storage and networking resources in a disaggregated way. But I think we’re now seeing the pendulum swing back, and 5G is going to enable an explosion of computing on the edge. What I mean by the edge is really anything from the customer premises or in the palm of the customer’s hand all the way to the central office of the telco. Right now we’re really limited in a lot of use cases that are possible at the edge because the connectivity is just not fast or reliable enough and doesn’t have enough coverage.

Within the 5G world, I would be investing a lot in anything related to edge IoT. There’s a lot of latent, industrial opportunity in what many would consider as relatively boring industries, boring in that they haven’t historically changed or evolved as quickly as those of us in the tech industry are used to. Farming and agriculture, for example, is a fantastic area that I think 5G will have an impact on. Natural resources are another area. The list goes on, and that’s where I would be betting big.

This is ironic because of the way cloud has often been mischaracterized. Because cloud people like things to be binary, there’s a perception that it’s either good or bad to be centralized. In reality we will see both continued massive growth in centralized computing resources as well as absolutely staggering growth of the number of devices, the diversity of devices, and the use cases on the edge of the network.

This is particularly exciting for telcos. There is enormous opportunity for innovation and growth in that conversation between the edge and the cloud, which telcos are intermediating. That will probably not be apparent until 5G is really made real. I know there’s a lot of folks saying 5G is here and that marketers are falling over themselves trying to advertise that we’re already in a 5G world. In terms of 5G technological adoption at reasonable scale, I’m thinking it’s probably another four or five years before we start to see that inflection point. I would definitely encourage the telcos to be ready to be that magical middle ground that intermediates.

ML — Do you have one last piece of advice for the telecoms listening to this podcast?

IR — I’ll probably bundle up multiple things in here. I’m going to repeat the know thyself, and I think it’s hard to know oneself. It takes a lot of introspection, work, and time.

I will also say understand the differentiated versus undifferentiated value and how that rolls up into your overall strategy. I think that sets up success in projects. Make sure that you’re applying your valuable developer resources on the right problems, while at the same time leveraging the commoditization of I.T. and the availability of free and open source software. I think that’s key.

Related to that, telcos are service providers, so know your customer. Working with enough telecommunications companies, I often find that we’re still steeped in traditional assumptions about who the customer is and what the customer wants. I think there’s a new generation of customers emerging in the business market who have new demands. I would be getting out there, putting a product management team, and really trying to understand how the customers’ needs are changing in the next five years.

It sounds a bit like being on a psychologist’s bench, but know oneself and know who you’re trying to serve and who you’re trying to delight. That is really fundamental. To me that means making more investment in product management and a new kind of product management. Traditional telco product management is so focused on analyst’s reports and what the vendors are selling.

You’ve seen this already in the web and e-commerce. It took a few decades, but with lean and agile approaches, we started to realize how dangerous it is to assume what the customer wants. Lean allows hypothesis-driven continuous testing of what the customer wants, which is helpful because what the customer says they want and what they actually want are often two different things. Without testing and talking with the market it’s impossible to figure out what customers actually want. You will find out more by engaging with the market as product management than you will in a vendor or analyst briefing. Market-centric user engagement is critical, and, in my view, will clarify the questions of differentiated versus undifferentiated heavy lifting, where the source provider should invest, and will lead to a strategy that allows the service provider to own their destiny, have its own differentiated strategy, and be less at the whims of analysts, vendors, and the general market.

At the end of the day, similar to I guess as an individual, you can really only own your destiny if you’re really taking full responsibility and accountability for it. I think what we’re seeing is getting out there in the market, testing the market, coming up with your own opinions and understanding of your customers is absolutely critical. And so that would be my very long but hopefully clear closing message for today.

ML — Thank you so much for your time and valuable insight. And it was really a pleasure to have you on this podcast.

IR — Thanks Maddison, I appreciate your time today.

This blog post was originally published here on CloudOps’ blog.

Sign up for CloudOps’ monthly newsletter to stay up to date with the latest DevOps and cloud native developments.

Leader in #cloud solutions, focused on open source, cloud platforms, networking, and DevOps. Experts in Kubernetes, OpenStack, CloudStack, and more.