Building a modern and open telecom network is never easy. Telcos and CSPs today face an innovation bind as each new generation of technology requires major capital investments which can be difficult to justify, at a time when competition is fierce and revenue outlook tough.
The traditional approach has often led to vendor-lock in, limited flexibility and difficulties in scaling. Building an open network is also complicated, with different hardware and software vendors touting their own solutions.
A quicker path for telco modernisation is being put forward by Dell, with the benefits of higher operational cost reduction, more agility and increased network efficiency. By involving partnerships with the tech ecosystem and leveraging a centralised IT approach, the new model also offers a structured pathway to leverage AI and faster ways to embrace new innovation.
To find out more about what Dell Technologies is doing to jumpstart the telco industry, the iTNews Asia team catches up exclusively with Dennis Hoffman, General Manager, Dell Technologies Telecom Business, during his recent stopover in Singapore, and discusses how the industry can get out of the current malaise of stagnant growth and how a cloud-based architecture and an open ecosystem can be the panacea for profitable and long-term industry growth.
iTnews Asia: Globally, many telcos are under pressure from declining profits. Why do you think this is so, and what do you see as the critical challenge the industry faces?
Hoffman: The industry is finding it hard to grow revenue – there's forced growth and costs. There's an investment cycle with every generation of mobile networking and technology. However, the majority of the revenue is coming from consumers who haven't seen significant benefits. Pricing hasn’t changed, while competition and expense growth remains significant.
For any business with flat revenues and forced cost growth, it's a tough challenge.
iTnews Asia: Has the time come now when telco players must change to remain competitive? How can cloud technologies help them?
Hoffman: Every country has at least one network operator that's a national treasure. Most governments won't let it go away. We can say the Telco industry must evolve. If COVID taught us anything, it's that network communications have become vital.
Almost everyone we're working with, including the world's largest network operators, are actively trying to figure out how to take advantage of what they’ve seen in the IT industry around cloud computing, cloud economics and cloud operating models.
They’re asking: How can we capitalise on that to become quicker and improve our total cost of ownership? How can we find a platform for new services and revenue streams, particularly from enterprises?
There is an imperative for telcos to modernise networks and find ways to better monetise them and create new services. To do this, they need an architecture that simplifies the process of launching new services while fostering collaboration across the ecosystem.
iTnews Asia: What do you see as the barriers preventing telcos from upgrading their networks and leveraging the cloud to achieve the outcomes you’ve outlined?
Hoffman: There is a high incentive (and pressure) on the network operator in the telco industry to never go down. It's such a vital service that the tendency is to avoid risk. Any industry that you know whose incentive system is dominated by mitigating risk will always be very careful about making changes. We can interpret that as a slow-moving and cautious industry because we need to regulate it, govern it and give it strong financial incentives to never go down.
Not surprisingly, the industry can't try every new technology. Telcos have to be convinced that it will grow their revenue, or will reduce their cost, and it has to be an appropriate level of risk to make business sense to change.
As a concept, the cloud is a type of computing requiring an architecture and an operating model. Leveraging both at the right scale impacts the economics. Architectures are also very vertical, and each new generation - 2G, 3G, 4G, and so on - requires major capital investments because the network is fundamentally not software upgradeable.
This cycle is visible in the market: during each upgrade phase, equipment suppliers thrive as operators make substantial investments. After that, operators enter a digestion phase, where suppliers see less demand until the next cycle begins.
What telcos increasingly want is a network architecture that allows them to make software-based upgrades, similar to how data centres operate. This would enable them to harness continuous innovations on general-purpose infrastructure without requiring a heavy investment each time.
This approach allows them to tap into the rapid innovation of general-purpose silicon, leveraging each cycle of Moore's Law. Instead of purchasing specialised hardware that depreciates over a decade and then needing to replace it entirely for the next generation, a cloud-based architecture runs on general-purpose infrastructure, meaning operators can continuously upgrade and benefit from new advancements.
iTnews Asia: What are your current priorities in working with the industry to address their pain points, and help support their transformation?
Hoffman: One major challenge with cloud-based digital transformation is its complexity. This transformation isn’t just technical; it includes an architectural shift, changes in the operating model, and even adaptations in their procurement and financial policies. Each of these demands careful attention, making the process a significant undertaking.
And anytime you make a change, it is multifaceted. It has that many parts and it’s going to be difficult. Our priorities are to help with all of that, to help them walk through what the architecture could look like and go beyond the architectural change with the operating model change, into areas such as procurement and finance.
The best way to take advantage of the general-purpose computing market is to stay on the leading edge of the processor curve, which is happening every three years.
Traditional investment cycles in telecoms involve large expenditures on each generation of technology, which is then depreciated over a decade. However, with general-purpose computing, staying competitive requires upgrading on a much shorter cycle - around every three years.
- Dennis Hoffman, General Manager, Dell Technologies Telecom Business
It means your competitor is running a network that is eight times as powerful with 50 per cent of the power consumption, and those two factors already change the economics dramatically.
Hardware subsystems are easy to upgrade. They're modular, which make them easy for an operations team to stay on the front end of the processor curve. We build software that facilitates that. By providing a unified ‘infrastructure block,’ we take responsibility for the entire cloud platform, helping operators reduce risk and simplify their transformation.
iTnews Asia: How does this transformation impact the decision-making process for telecom operators and their collaboration with industry partners like yourself?
Hoffman: In the current model - which is still dominant, rather than outdated - telcos often assign specific geographic segments of their network to particular vendors. For example, in the United States, an operator might assign Manhattan to Nokia and Los Angeles to Ericsson, creating accountability for each vendor within their assigned region. This way, the performance of the network in a given area becomes the responsibility of one supplier, allowing operators to manage competition by varying the geographic distribution awarded to each vendor.
In contrast, a disaggregated cloud model introduces multiple vendors for different parts of the network. One vendor might handle the applications, another the virtualisation software, another the hardware, and possibly even a different vendor for the radio access network components.
This complexity raises a key concern: if there’s an outage, who is responsible for identifying and fixing the issue? Immediate network restoration is critical, and in the traditional model, accountability is clear-cut. In a cloud-based model, some fear a lack of accountability, potentially high systems integration costs, and difficulties in coordinating issue resolution.
Although this concern is common, it’s worth noting that in a cloud model, there is generally less need for deep systems integration between components and more focus on validation and certification.
In an industry that is generally risk-averse, such sweeping changes can be challenging to implement. Moving to this new model brings major changes to network architecture, operations, procurement, and financing. A recent Merit Talk study found while 96 percent of telcos believe that a cloud architecture and network disaggregation are the way forward, but not many have started making this transition yet.
iTnews Asia: Given their reservations, how then do you convince telcos to move beyond the traditional, risk-averse model? Is it more a matter of changing their mindset, or helping them see how they can integrate different innovations and work effectively with new partners?
Hoffman: I don't think we’re ever going to fundamentally change the risk appetite. Instead, it’s on us as an ecosystem to deliver solutions that reduce the perceived risk. To do this, we need to collaborate effectively to provide comprehensive answers. Partnerships are crucial in achieving this; it’s about seeing ourselves as an ecosystem focused on delivering reliable, integrated solutions.
We have deep partnerships across the ecosystem, with players like Dell, Ericsson, Nokia, Samsung, Red Hat, VMware, and Wind River. By ensuring that our products work together seamlessly and supporting operators with unified accountability, we can offer an open architecture with a 'single throat to choke - one point of contact who can take responsibility if issues arise.
This approach also requires a holistic view of network architecture. We need to work with the industry to make strategic decisions about which technologies to use and where they are best applied. Then, we implement these solutions in a focused area first, testing and refining until we’re confident with the model before scaling up. This is how global leaders are effectively tackling these challenges.
Most start with the areas of highest capital intensity, like cell towers, where every deployment represents a significant investment. For instance, the core network might involve just two national data centres operating redundantly, which closely resembles the IT model - centralised and manageable.
But there’s a fundamental difference between managing thousands of servers in two data centers versus managing a single server in 5,000 individual cell towers. The operational complexity of distributed cell towers is vastly greater.
What sets this model apart is the unique demands of cell towers, particularly the radio equipment, which is both costly and power-intensive. It’s these specific challenges that drive a more tailored approach to network architecture.
iTnews Asia: Could you share some key highlights about your ecosystem partnerships? Are there any notable collaborations or initiatives you’re working on with partners, especially regarding the labs and how they align with your overall strategy?
Hoffman: Yes, OTEL our open telecom ecosystem lab, is an integral part of our strategy. To address some of the key challenges, we saw a significant gap in the market—no one was building hardware specifically for telco networks.
Hardware for radio access networks needs unique features: a specific depth, directed airflow, front-facing serviceability, and compliance with strict standards. It also must operate reliably across extreme temperatures, from freezing conditions to desert heat. Since no one else was addressing this, we stepped in to develop the necessary hardware.
Another critical challenge was ecosystem compatibility. We need various components to work seamlessly together, which requires ongoing testing. We invite ecosystem partners to participate, assemble the components, test them to ensure compatibility, and keep testing to maintain this over time. Ensuring continuous compatibility is a much more complex task than a one-time setup.
Key partners include those at every level of the stack, especially silicon providers like Nvidia and Intel. In total, we collaborate with around 120 partners.
iTnews Asia: What is your AI strategy, and how does it align with the findings from the recent study, which indicates that while many see AI as essential for transformation, they also struggle to keep pace with its implementation?
Hoffman: It's crucial to clarify what we mean by AI, as it is a broad and often overhyped concept. We use a simple framework that defines AI as a tool or capability. The key question is: what are you using it for? AI can be applied within telecommunications networks to enhance performance -making them better, faster, cheaper, or less risky. However, there are also risks associated with its implementation.
In the network context, AI helps network operators leverage the technology to generate revenue by delivering AI solutions to end users. Similarly, telcos can adopt AI in their organisations to improve operations, like customer service, much like any other industry such as banking.
Focusing on AI in the network, there is already an ecosystem of software companies providing telco-specific AI capabilities. We have partnered with several of these companies, recently announcing our collaboration through the ‘AI for Telecom’ initiative. These partnerships have emerged from ongoing work with operators worldwide, where we have encountered these companies during trials and proofs of concept.
Digital transformation encompasses a wide range of challenges, and the question becomes how we can assist operators. We advise them to start by narrowing their focus. Instead of discussing AI in general, telcos should consider specific applications, like using AI to improve their networks or grow revenue by serving enterprise customers.
Two fundamental principles guide our approach: first, bring AI to data rather than moving data to AI, as it's more efficient to place infrastructure near data sources. Second, bad data leads to poor AI outcomes, so having a robust data strategy is essential. This includes ensuring data quality and understanding the data pipeline from storage to processing.
Once operators have defined their use cases and established their data strategies, we can discuss the necessary infrastructure and architectural considerations. This includes selecting the appropriate AI model - whether it's a large language model, a medium model, or another type tailored to their needs.
iTnews Asia: How do you envision the journey of building the telco ecosystem with your partners, what aspirations do you have for its success?
Hoffman: Currently, the industry is transitioning through the initial rollout of 5G and is not yet considering 6G, creating a unique opportunity for network operators to modernise their architectures. It is essential for us to assist telco leaders in this modernisation process before the onset of 6G, as their advancements could set a standard for the rest of the world.
However, this transformation won't happen overnight; it will take the entire 5G cycle, with much of the implementation occurring during the 6G phase. We recognise that the biggest challenge lies in the transformation itself, which involves both technological and architectural changes, as well as shifts in workforce skills and job descriptions.
While some operators are eager to lead the way in innovation, others are hesitant and prefer to follow once they see success. Our role is to identify and support those willing to take risks while preparing for the subsequent wave of adopters.
Looking ahead five years, the digital maturity of telcos will likely reflect a bell curve of progress, with many evaluating their options, a few leading the charge, and some lagging behind. Factors such as average revenue per user (ARPU) and market competition will significantly influence this evolution, resulting in varied maturity levels.
