‘So what’ and disaggregation spell success in Coordination Age

BT Group’s Gabriela Styf Sjöman talks to Contributing Editor Annie Turner about her approach to harnessing technology and why network readiness is all. 

Gabriela Styf Sjöman joined BT Group as Managing Director, Research and Networks Strategy, in June 2023. Previously she was Group Chief Strategy Officer at Nokia and has held senior roles at Telia, TIM and Ericsson as well as sitting on various boards as a non-executive director.

How does Styf Sjöman sum up her role? “To create an awareness and influence about the opportunities of new technology,” but, she stresses, always concentrating on the “So what?”. In other words, what technology can do, not for its own sake.

She explains, “It’s always technology that has driven new innovation…for instance, a big shift was the separation between consumer service and network that came through IP. Then we were all taken by surprise to have all these services going over the top. We didn’t ask ourselves, ‘So what?’ about IP.”

Styf Sjöman firmly believes that, “If we really want to understand that ‘So what?’ in society and [for] societal growth, economic growth, then we need to have that multi-disciplinary understanding”. Her education and career are multi-disciplinary.  She originally studied power engineering before going into telecoms, with her first job at Ericsson. Among other things, she has studied law and political science, and holds both an MBA and MA in international affairs with a specialism in cybersecurity.

Multi-disciplinary doesn’t mean lacking clarity: “I’ve led R&D before, overseen a lot of innovation. I always say, you are either looking to make money or save it. There’s nothing in between. You must be able to explain your value proposition to your customer,” she says.

The next big step

She sees disaggregation as the next big network evolution. “Cloudification of the network is not only for optimisation and to reach hyperscale economics. Disaggregation is opening value chains. That’s why we see new, different players coming into connectivity and what the full stack of an integrated telco used to offer. We have new players on services, on the platform level and in infrastructure – netcos and towercos,” Styf Sjöman says.

She adds, “APIs are a big ‘So what?’ – abstracting network capabilities and opening them to developers and customers – and becoming AI-driven for self-optimisation, etc. All of this is transforming the network to become a platform that enables product innovation on top of it and we need to treat it as such.

“We should be asking ourselves, what does this all mean for the role of the network? What does the network become? In my eyes, the network becomes a production environment – a programmable network that enables the fast development, introduction and operation of new products, either for myself or the ecosystem. I open up my network to monetise the ecosystem and unleash new product innovation.”

Too slow, too expensive

Styf Sjöman says she only understood why telcos have so few products after she went to work for one. She discovered it cost perhaps tens of millions of pounds and took two years to ready the network and associated IT for one new product and its launch. She says, “Now we’re creating all these disaggregated, more agile, faster networks. The networks must serve product innovation, allowing me to develop and introduce even remove applications at a much faster pace to create that long tail of applications.”

She describes the industry as coming to the end of the Information Age, in which people and things are connected to share information, and the beginning of the Coordination Age. Here “collaboration and coordination are what matters, so everything we connect is coordinated: AI, machine learning and analytics are for actionability.”

Characteristically she asks, “What’s the role of the telco?”. Clearly, connectivity is the base for it all, but “what are the new capabilities I need? What are those business models?”.

She notes the shortcomings of received wisdom. For instance, while everybody talks about the need for software skills, Styf Sjöman thinks software architecture is the critical issue. She notes, “Increasingly we buy everything off the shelf but as we move towards programmable networks, parts of the network will be unique to specific telcos. We will need our own system integration and development capabilities.

“It’s not only developing a few lines of code…You need to know how to architect a product, how to develop and maintain it, and to know how to do the overall product management. It’s about new functional skills. This is usually described as the journey from telco to techco.”

An eye on the horizon

Then there’s the importance of keeping an eye on the horizon. For example, BT Research is working to understand the impact quantum computing will have on quantum networks. Styf Sjöman states, “A quantum computer will be able to decrypt any cryptography in milliseconds and it will come. Our research has been around quantum secure communications, such as quantum key distribution.”

The idea is to understand how the nature of quantum can also be used to protect different type of keys and cryptography. BT (working with Toshiba) has scored a world first in quantum secure communications, having launched a trial commercial quantum secured metro network in London. EY was the first customer to join the network, followed by HSBC which has become the first bank to pioneer quantum protection for AI-powered foreign exchange trading, using the network.

No wonder Styf Sjöman thinks we’ve only glimpsed the tip of the iceberg regarding use cases in the imminent Coordination Age. They will be powered by automation, from autonomous manufacturing to using drones for public safety, in a remote health care and much more. Their common denominator is that they will require an intent-based network.

Intent runs deep on automation and AI

She defines this as a network that is service- and device-aware that is on-demand with seamless access regardless of access technology. “So everything is around very deep AI that scales in [step] with self-optimised networks,” Styf Sjöman explains. “There will be an absolutely massive shift in how we operate the networks and assurance in future. We will also require skills, process redesign and automation of those processes.”

She says at the moment, telcos are using AI within domains and on a component level, rather than end to end. She emphasises that in future, “If the network is ‘going to serve’, there will not be ‘a killer service’, but thousands and thousands of applications. Some will continue over the top, others will be net-compute applications.”

They will require a much higher level of interdependency between the consumers’ cloud-native applications and the network application. This means seamless optimisation and management from the core network all the way to the radio. Furthermore, the thousands of net-compute apps will require network slices.

Styf Sjöman asks, “How are you going to get economics out of that? You need an AI-driven network that understands the service and the device to decide which is the optimal performance to deliver the best customer experience. So where do I put my workloads? Which infrastructure do I use to deliver that performance and the best economics, sustainability and energy consumption? That’s that level of AI you need.”

How’s it going?

How much of this is theory and how much actual? “We’re progressing a lot at BT, such as our global fabric, the product that we launched for BT Business [see this recent interview with CTO of BT Business, Colin Bannon]. It is like the first generation of Network-as-a-Service [NaaS]. It allows BT’s customers to request connectivity like Ethernet links and broadband access on-demand, leveraging the benefits of virtualisation to instantiate and activate network functions on-demand.

“The next step is AI-driven self-optimisation for an intent-based network. Nobody has it yet but that must be the objective and disaggregation is key.”

Styf Sjöman says making the most of those opportunities will demand far more concentration on product innovation: “We often introduce innovative network technology, but mostly use it to sell the very same products. This needs to change. We need more focus on exploring how we monetise the innovation to find new business models, in collaboration with our customers and partners.”

She notes, “Customers don’t wake up and say, ‘I want a private network’ or ‘a 5G slice’. They’re trying to solve for an outcome, to optimise their business. They are thinking of B2B2X business. They don’t only want to use our network for their digitalisation but also to build on top of our networks to offer products to their customers.”

Integration and differentiation

She concludes, “We have only seen the tip of the iceberg of the digital services that will underpin societal and economic growth in the UK. Many of these services will remain over-the-top and be network agnostic. But there will be a set of services that are so critical that they will require the best a carrier grade network has to offer in terms of, for instance, security, speed, latency and jitter – and delivered as a cloud delivery model.

“To be integral as an enabler in the Coordination Age value chain, the network must be intent-driven, self-optimised, disaggregated, and open. And this is exactly what our refreshed BT network vision is about.”

Interview – Telco-Techco: Capturing the Opportunity – Steven Stephen Spellicy, VMware at FutureNet Asia 2023

Recorded at FutureNet Asia on the 19th of October 2023.

1. What are the business models and how do telcos translate the models into actual services and monetise?
2. We have been talking about the telco-techco transformation for some time, what is the state of play and how much progress has been made to date?
3. How will ecosystems drive growth in future for telcos and how will telcos fit into the new ecosystems?
4. What are the operational challenges for telcos on this journey from telco-techco and how can automation be leveraged to accelerate the shift?

Interview – Steven Zheng, Huawei at FutureNet Asia 2023

Recorded at FutureNet Asia on the 19th of October 2023.

Huawei’s Autonomous Driving Network Strategy.

Interview – Sam Keys-Toyer, Ericsson & Andrew Collinson, Connective Insight at FutureNet Asia 2023

Recorded at FutureNet Asia on the 19th of October 2023

Interview – Vinod Joseph, VMware & Lian Jye Su, Omdia at FutureNet Asia 2023

Recorded at FutureNet Asia on the 19th of October 2023.

Interview – Niall Norton, Amdocs & Roy Chua, AvidThink at FutureNet World 2023

Recorded at FutureNet World on the 4th of May 2023.

Interview – Susan White, Netcracker & Roy Chua, AvidThink at FutureNet World 2023

Recorded at FutureNet World on the 4th of May 2023.

FutureNet World Speaker Interview with Stephen Spellicy, VMware

Following FutureNet World (3/4 May, London) we caught up with Stephen Spellicy, Vice President Service Provider Marketing, Enablement and Business Development, VMware to get his thoughts on some of the key questions raised at the conference.

How can we accelerate network modernisation programmes and what are the biggest obstacles?

At a technical level, standards and reference architectures have made huge inroads when it comes to CSPs’ ability to innovate and drive down cost, but this only gets us so far. While VMware has a number of programs to enable technology partners and ease interoperability across the core, RAN and cloud, CSPs are looking to our automation and assurance solutions to de-risk their multi-vendor environments and accelerate deployments in the field. By leveraging automation and assurance to re-architect service operations, the industry is reducing bespoke integration efforts and reducing their OPEX. Beyond the cost savings, this shift in operational approaches will lead to faster, easier paths to differentiate-able innovation.

Re-architecting operations is not an easy challenge to solve. As highlighted by several CSP panel sessions at FutureNet World however, it’s an area of focus that could shift the industry’s trajectory. To get there, CSPs need to join forces with members across the industry to help facilitate industry-wide engagement focused on simplifying network operations.

As operational models shift, the next logical challenge is the skills gap. Another topic well covered by the sessions at the event, many CSPs feel that while we’ve witnessed tremendous progress over that last ten years since NFV was introduced, there is still concern that the ‘right’ skills to run these new operations are not yet available inside CSP organizations. As architectures have been shifting from siloes to horizontal approaches, one approach to filling the skills gap is to flatten the operational structure; remove service silos and replace them with horizontal teams. Hardware platform, cloud platform, and applications may be more in the domain of traditional IT skills whereas network operations, service orchestration and regulatory observance may be more suited to traditional network ops teams. By building layers of teams that span all services, I think the skills issue can be re-defined and the gap may not be as great as some perceive it to be. Standardization of technology platforms greatly reduces the issue of skills gaps, as organizations embrace and invest in platforms, training, enablement, and knowledge transfer becomes much easier. VMware provides several certification tracks to enable IT and network professionals with training, certification and tools to simplify the adoption of our platform. As an industry leading platform, there are also user groups and online communities that can be leveraged to share best practices to make the best use of our platform service, both within the Telco domain and in enterprise IT.

How can we overcome obstacles like investment levels to speed progress?

Industry investors see the telco industry as a relatively stable, low margin business returning modest but perhaps ‘safer’ return over time than other investments. Many CSPs have made huge investments in recent years to build and deploy 4G/5G networks, which has created greater pressure to show a return before significant future investment can be sought. To accomplish this, the technology industry is adopting OPEX-based (subscription and SaaS) sales models for their CSPs sales. While CSP have traditionally used an OPEX-based model on their revenue side; selling services on a usage and subscription basis, they typically view infrastructure as capital expenditure. This causes a disconnect in the way CSPs view and are viewed financially, based on how their budgets are defined and accounting models are typically constructed. As an industry insider, I would like to see a shift in perception of CSPs as a more dynamic and agile industry, regarding both their operational and procurement models. The benefits of a consumption model can be seen across IT and network domains for the communications industry, as many of the leading software and hardware providers are already driving this transformation in key areas of the CSP operations, such as OSS/BSS, customer care, CRM, and NOC/SOC tooling.

Watch Stephen’s panel discussion at FutureNet World 2023

How can we leverage automation to implement new partnership models and overcome the challenge of competitors collaborating on projects?

Importantly, automation facilitates collaboration and eases integration challenges among vendors as we jointly develop new service models for CSPs. This co-innovation often requires a complex integration of different technologies, applications, and hardware—an often time-consuming and error-prone process, especially when multiple vendors are involved. Network automation can address these challenges by streamlining the entire service delivery process, from infrastructure provisioning to network configuration and service deployment. When these tasks are automated, partners can reduce the time, resources and financial obligations required for collaborative service delivery, while also reducing the risk of errors and delays when building a cohesive offering. VMware, and its focus on centralized automation, helps CSPs to standardize service delivery across multiple vendors, improving consistency and reducing the risk of service disruption. As such, this helps ensure that CSPs can create high-quality services that meet (and exceed) their end users’ requirements.

When do you think CSPs will really scale AI and automation, moving beyond applying it to specific tasks to wider capabilities and end-to-end processes?

Today, many operators are leveraging basic forms of AI and combining it with automation to drive specific operational tasks. As both technology areas evolve, the solutions will become easier to use and enable more advanced feature sets. This is a natural evolution for AI, as the systems are trained through experience. One key area we will soon see the acceleration of new capabilities enabled by AI, is in the RAN. As we move to more open and agile RAN, leveraging AI-driven intelligence will be essential to tapping into unified network data and programming the network to instrument change based on that data. The RAN Intelligent Controller (RIC) is where this already coming to fruition.

FutureNet World Speaker Interview with Sam Keys-Toyer, Ericsson

In the leadup to FutureNet World (3 & 4 May, London) we caught up with Sam Keys-Toyer, Head of Business and Portfolio Development, Managed Services Networks, Ericsson to get his thoughts on how CSPs can overcome operational complexity, the barriers to operational transformation programmes, the need redesign operations to support AI, and much more.

 

  1. Organisation, structure and culture: why is this still a barrier to driving through operational transformation programmes?

[Sam Keys-Toyer] It is certainly true that a key to successful transformation is in the hearts and minds of the people affected. A new functional and organization model is required to enable this digital transformation to service-centric and data-driven operations. If they embrace the vision and journey the chances of success increase exponentially. On that basis a smart change program is paramount supported by strong and active communications, with visible leadership involvement that engage people to propel the transformation forward.

 Transformation cannot be solely focused on amending what is happening today, the real challenge lies in how to get there – it must be based on a clear vision on what the organization wants to achieve and deliver in the future. Defining this clear vision anchored on tangible and measurable business outcomes and a solid investment plan, supported by executive sponsorship must be the starting point.

Also, cultural norms for how people collaborate, especially between functions or teams, must evolve. A proactive mindset and cross domain working become default behaviors. There is a need to promote agility, speed up decision making, minimize bureaucracy and liberate peoples’ problem-solving capability.

Ericsson are a Sponsor of FutureNet World, 3 & 4 May, London

When we analyzed our functional structures, accountabilities, governance, skills & behaviors, and ways of working, we revealed significant gaps between our needed culture and the reality on the ground, not only necessitating a redesign of the people blueprint to a flexible model but to also execute a very effective organizational change program. Big bang approaches typically don’t work, especially when technology adoption is part of the mix.

 

  1. There is a clear need to build intelligent operations for a 5G world: do you think there is a need to redesign operations to support AI?

[Sam Keys-Toyer] As 5G becomes more common and eventually 6G is introduced, network management will become more complex and require greater agility and minimal manual intervention to ensure complexity is simplified for operations. It’s important however to have the right focus. The goal of implementing AI techniques in network operations is not for the technology itself but to deliver tangible business benefits that CSPs will be able to extract only if they consider a number of keys parts of any operating model redesign or transformation, and which from experience should be addressed:

Ericsson´s holistic approach to transformation

Ericsson´s holistic approach to transformation

It is important though not to be distracted by the “shiny new things”. While key enablers, automation, analytics & AI are techniques or enablers and not the solution to digitalization in isolation. Business value must be front and center to all automation and AI ideation, development, and deployment.

 

  1. In order to unlock the 5G opportunities in both enterprise and consumer markets, CSPs need to overcome operational complexity at speed: How can they accelerate the adoption of new operating models to leverage these opportunities?

[Sam Keys-Toyer] The next evolution of the Operating models must support a self-optimizing network driven by intent and underpinned by hyper-automation. This involves transitioning from people managing the network to people managing the machines that manage the network. For this to work, data/knowledge, policy, automation, assurance, analytics, machine learning and reasoning, and security must be integrated into a true Intent-Based Network.

This in turn will enable a wider variety of applications and use cases for consumers and businesses with CSPs aiming to provide configurable services with detailed agreements on functional and non-functional characteristics that require dynamic adaptation to the network. With this increased flexibility and dynamic adaptation, the network must be managed in real-time to deliver on these requirements. Concepts like “autonomous networks” and “intent” are needed to drive new propositions and automate the network’s state to meet performance KPIs, SLAs, and business outcomes.

We see three main drivers for intent in building and operating new services:

  1. The complexity and cost of operating 5G networks and beyond require a new level of automation – a level beyond even AI/ML today’s automation
  2. The need to transform operations to better meet the needs of the business
  3. The need for transparency and AI explainability – so that we can trace back all decisions recommended and actuated on by the system.

 

  1. The industry is well aligned on the need to transform operations to better meet the needs of the business: can you give some indication to the tangible business benefits gained by the adoption of automation and AI?

[Sam Keys-Toyer] Automation & Artificial Intelligence are essential to get the most from our data driven operations approach. The massive scale we manage enables global feedback loops that we leverage to constantly evolve our processes, grow our closed-loop automation & AI use case libraries, making it possible to handle a global network of more than 710 thousand sites and 6.4 M RAN cells with highly focused human intervention. To illustrate, the aggregated level of close loop automation within our today’s network operations is above 88% and in 2022 we reached 36 M of AI recommendations.

This has a significant tangible impact on CSPs operations business outcomes. From a network performance perspective, data-driven operations reduced network unavailability by 34 percent while decreasing customer complaints by 21 percent. On the network efficiency side, the transformation led to a significant 12 percent reduction of work orders (WO) and 24 percent less truck rolls per node and up to 8% reduction in energy consumption which has an important impact on CO2 emissions.

FutureNet MENA Speaker Interview with Sam Keys-Toyer, Ericsson

In the leadup to FutureNet Middle East & North Africa (20 March, Dubai) we caught up with Sam Keys-Toyer, Head of Business and Portfolio Development, Managed Services Networks, Ericsson to get his thoughts on Ericsson’s strategy and the topic of Intent-based network operations: unlocking the potential of 5G monetization.

5G continues to advance at incredible speed unlocking new monetization opportunities across enterprise & consumer segments. However, as the possibilities multiply, so does the complexity.  How are you addressing this complexity through data-driven operations?

Absolutely 5G is growing fast, particularly 5G SA which is also accelerating and it is the real foundation for e2e network slicing that allows for differentiated performance down to the device level. When we look under the hood of many Operations today, we see that virtualization is driving notable complexity starting from the Core Network transformation to a fully cloud-native environment, often as the first step to introduce 5G Core.

We know that adding people around legacy processes is not sufficient to deliver the needed volume & speed of change. A new look at data, processes & automation is key to turn operations into data-driven. Today we have 100 managed services contracts globally, out of which 20 have been transformed to Ericsson Operations Engine, which is a multi-vendor, multi-technology framework, basically the heart of our data-driven operations approach. It cuts across multiple dimensions, from processes through competence, organization, strategy, governance, data, automation & artificial intelligence (AI).

What are the actual tangible benefits of those 20 Operational transformations you mentioned and how did the adoption of automation and AI help better meet the needs of the business?

We currently manage an unparallel multinational, multi-vendor digital data store that serves 1 b subscribers across 100 countries.  This massive scale enables global feedback loops that we leverage to constantly evolve our processes, and grow our closed-loop automation & AI use case libraries, making it possible to handle 18.2 M work orders and 5b alarms every year with highly focused human intervention.

This has a significant tangible impact on CSPs’ operations business outcomes. From a network performance perspective, data-driven operations reduce network unavailability by 34 percent while decreasing customer complaints by 21 percent. On the network efficiency side, the transformation led to a significant 12 percent reduction of work orders (WO) and 24 percent less truck rolls per node and up to 8% reduction in energy consumption which has an important impact on CO2 emissions.

Automation & artificial intelligence are essential to get the most from our data-driven operations approach. Not just for the shake of the technology itself but to maintain the highest network performance and customer experience standards. To illustrate, the aggregated level of close loop automation within today’s network operations is above 88% and in 2022 we reached 4.8 M AI-triggered actions, so yes automation and AI are key in the transformation process but they are not the only ingredient. And in fact, there are other foundational aspects that need to be established before for AI to deliver value, such as process transformation, a futureproof application landscape, or a solid data strategy.

So, why do we need to shift towards data-driven intent-based operations?

Let’s start with what we mean by intent-based. Intent is basically a statement of what we want the network to do or the outcome we want from the network. It is not specifying how the network should achieve that goal, which could be by delivering on stringent SLAs or improving the net promoter score, or reducing energy consumption. We see lots of examples of autonomous intent in everyday life if you use virtual assistants such as  Siri, Alexa or Google Assistant, but also we should remember we have been using Intent statements for years as Humans – “ I want my network to have the best coverage, fastest speeds and lowest cost “.

To monetize the 5G investments, the industry will want to deploy more and more diverse services and open up new customer markets. The network will need to support all such cases and adding more humans to cope with this demand will not scale.

We see three main drivers for intent in building and operating new services:

  1. The complexity and cost of operating 5G networks and beyond require a new level of automation – a level beyond even AI/ML todays automation
  2. The need to transform operations to better meet the needs of the business
  3. The need for transparency and AI explainability – so that we can trace back all decisions recommended and actuated on by the system.

Where are we on the development of Intent-Based Operations?

The current stage of industrialization is taking us from the test network to the live 5G networks of our customers. The purpose is to experiment with real network data and real slices in order to strengthen the core capabilities of the system.

Alongside the core capabilities, a system that is replacing human cognitive effort has also to win the battle for trust. Where we are paying special attention to explainable AI.

So in summary our research and investment into Intent Based Operations continues. As we prove out the industrialization activities, we will be widening the scope for more field trials during the year before making the system generally available.

Intent-based operations is an inevitable shift in the approach to deploying and operating 5G services, leading us to a more autonomous and cost-effective approach.  However, I would also like to leave you with the question: Why not apply the same philosophy to 4G services, even more opportunities for efficiency?

 

If this article and topic interest you, we recommend you read more here: Data-driven operations. Does it pay off? – Ericsson


Warning: Invalid argument supplied for foreach() in /home/customer/www/futurenetworld.net/public_html/wp-content/themes/futurenet-group/template-parts/content.php on line 13

Interview with Basar Akpinar, P.I. Works at FutureNet Asia

Recorded at FutureNet Asia on the 19th of October 2022.