CTO Technology Trends 2024: Key trends shaping the future of digitalization

In the 2019 edition of the CTO tech trends article, I outlined my vision for the network platform of the future, emphasizing the crucial role of consistent and open business interfaces – that is, application programming interfaces (APIs) – that offer a multitude of capabilities to all users. Since then, we have catalyzed a shift in the industry, placing this vision at the forefront not only in the mobile sector but also among partners, enterprises and developers. 

As a result, we are now on the cusp of a massive transformation as the telecom industry transitions toward a platform economy. The mobile industry is today dominated by smartphones and best-effort mobile broadband (MBB) services in a mature market, and service innovation is a top priority on the industry agenda to spur profitable growth by utilizing new business models and more advanced network characteristics.

As we embark on this transformative journey, our overarching objectives remain clear: to ensure the cost efficiency, sustainability and trustworthiness of high-performing programmable networks. Ericsson has prepared a toolbox, equipped with a developer platform, network APIs, network exposure and network slicing to enable differentiated services. Equipped with these powerful tools, we stand ready to shape the next era of wireless connectivity and business innovation.

Analyzing the evolving needs of society and business

Both the business community and society at large are continuously searching for more efficient and sustainable solutions with the potential to improve profitability, provide economic stability and stimulate growth. Today, many of these solutions are enabled by digitalization, electrification and automation. The electrification of everything includes the energy transition toward renewable energy sources such as solar cells, distributed energy storage and the transition from combustion to electrical engines. Advances in automation include autonomous operations and artificial intelligence (AI) powered learning, reasoning, problem-solving and language understanding, with generative AI playing a significant role. Generative AI not only boosts productivity in development processes but also plays a pivotal role in fostering innovation. Examples of this include efficient data analysis that provides insights and predictions, the personalization of experiences and new methods of delivering products and services.

The digitalization of society and businesses is advancing rapidly in tandem with the evolution of connectivity, AI, compute and cloud-infrastructure technologies. In the case of 5G networks, these technologies are used to facilitate programmable and differentiated networks for more efficient service delivery. The performance, integrity and versatility of the global 5G network serve as a powerful digitalization catalyst.

Ericsson’s mission is to lead our industry in creating a stronger, more vibrant and more innovative business landscape. We are also driving cross-industry collaboration to form an open network ecosystem with leading industry players from devices and chipsets to the cloud providers and application developers. The global network platform we envision will be easily accessible to any enterprise or developer, and will address society and business needs through viable business solutions that spur innovation, collaboration, competition and investments.

The challenge of meeting differentiated needs

Today’s consumer, enterprise and developer applications – as well as those of the future – comprise a vast array of use cases with very different network needs. These network needs range from low resource-utilization requirements, such as text messaging, to more resource-intensive needs, like immersive communication experiences. Requirements on throughput, latency and other network characteristics to optimize the user experience vary widely.

For instance, consider the contrast between video streaming and immersive communication. Video streaming employs buffering to ensure a smooth user experience, accommodating potential network fluctuations. On the other hand, immersive communication relies on near-real-time interactivity to provide a seamless experience, allowing users to interact with virtual objects, physical environments and other users.

In addition to variations in terms of network resource demands, different applications and use cases also have widely varying requirements when it comes to reliability, coverage, security, identification and authentication. For example:

• Consistent and reliable connectivity is crucial for government services, such as public safety operations.
• Specific uplink characteristics are necessary for interactive immersive communication.
• Enterprise control over all assets, achieved through autonomous identification and authentication, as well as a secure access service edge, is imperative.
• Comprehensive outdoor and indoor coverage, along with advanced positioning capabilities, is essential for enabling features like 3D mapping support and location checks for fraud protection.

The high-performing differentiated network will address these various needs in a cost-effective and resource-efficient manner. Given that high performance is not always a requisite, optimal efficiency can be achieved through the use of orchestration techniques that dynamically adjust network resources to meet actual needs, while simultaneously ensuring the required quality of service (QoS) for the different user groups. This underscores the importance of future networks being not only open but also programmable.

The high-performing differentiated network approach caters to new segmented and dynamic business models spanning application, device, network and cloud. While today’s besteffort subscription will continue as a basic service offering, the differentiated approach will enhance the value of network offerings and subsequently increase the willingness to pay.

Four trends driving the evolution of high-performing networks

The pace of mobile innovation over the past 30 years has been extraordinary. For instance, we have achieved a 100-fold increase in data speed every decade, amounting to a 1 million-fold leap over 30 years – surpassing the growth projected by Moore’s law. In the next decade, we anticipate another 100-fold improvement. This will be driven by the combination of high-performing programmable networks, service differentiation underpinned by AI-enhanced network automation and groundbreaking technological innovations in both software and hardware. Together, these elements will lead to a new era of connectivity and business innovation.

TREND #1: Using network programmability to meet the need for service differentiation

Network programmability fundamentally revolves around two key aspects. The first is the programmability of the network infrastructure, which is managed by the communication service provider (CSP) to efficiently utilize and operate the network resources. The second aspect is on-demand service programmability, which allows applications to dynamically request network resources using network APIs for various service levels. The network APIs are accessible through a network infrastructure exposure layer to network aggregators, such as Vonage, which offers globally available network APIs and services to the ecosystem of application service providers and application developers.

The CSP exercises control over the network resources, verifying availability. This process is facilitated by the management system of the programmable network, which employs AI-based observability agents and mechanisms. These agents enable the system to automatically respond in real time as well as anticipating future needs when required, thereby ensuring optimal service provisioning and network operations utilization.

The interplay of network infrastructure programmability and network APIs is fundamental in catering to applications and services that require both diverse and dynamic network performance metrics. Hence, programmable networks will offer predefined performance characteristics that range from best effort to guaranteed, thereby meeting the array of connectivity needs. These predefined performance characteristics can be structured into four main classes:

• Class 1: Adaptive Buffered – best-effort throughput and best-effort latency that caters for applications developed for traditional MBB.
• Class 2: Fixed Buffered – guaranteed throughput and best-effort latency optimized for use cases like background application communication, such as over-the-air upgrades.
• Class 3: Fixed Immediate – guaranteed throughput and guaranteed latency optimized for applications that demand high reliability and predictability, such as live broadcasting or the remote control of vehicles.
• Class 4: Adaptive Immediate – best-effort throughput and guaranteed latency optimized for applications that require consistent response times, like uninterrupted live video feeds or video calls with adaptable resolution.

Guaranteed performance is subject to coverage and can be provided with different levels of performance metrics, offering further degrees of differentiation and customization based on specific application needs. The programmable network infrastructure will therefore provide advanced observability with high granularity and dynamism through mechanisms such as network slicing.

Gaming is an excellent example of a use case that utilizes all four performance-characteristic classes. Class 1 is sufficient for tasks such as downloading menus, while class 2 is best suited to receiving new maps/terrains, for example. Class 3 is the optimal choice for actions in competitive multiplayer games, while class 4 is ideal for tasks such as receiving remote-rendered screens.

The required guaranteed performance levels are requested by the applications through the network APIs, which enables the CSP to charge for the guaranteed network performance provided. The guaranteed performance can be purchased in different ways, such as a one-time purchase triggered by the application or a consumption-based purchase bundled with the application.

In summary, programmable networks enable service differentiation for consumers and enterprises by providing guaranteed QoS. This not only ensures robust and reliable connectivity but also empowers CSPs to monetize network services and capabilities, including Service Level Agreement assurance, observability and network APIs. This will enable CSPs to compete on both customization and performance.

TREND #2: Leveraging telecom AI to boost the programmability of the high-performing network

The programmability of the high-performing network will enable its evolution into an intent-driven, autonomous system, capable of supporting diverse service needs with optimum cost efficiency and sustainability. This transformation will be powered by AI, which will drive performance optimizations across business, service and resource network operations. The solution entails a hierarchical structure of embedded AI in core networks (CNs) and radio access networks (RANs), AI-based management and orchestration functions and generative AI-based network operation. This paves the way for autonomous network operations and optimization of capabilities such as energy performance, capacity, coverage, throughput and resilience. Humans maintain control, providing system requirements and overseeing their fulfillment, by expressing intents rather than dictating specific actions.

AI-powered execution environments must be available everywhere, and AI training could be colocated if needed to support the increasing usage of data for multiple purposes. The distributed data infrastructure is managed through a common data plane to enable efficient life-cycle management, transport and storage of network data. The life-cycle management relies on a fully automated trustworthy model that constantly improves and follows data changes, to achieve system-wide end-to-end gains, such as constantly improved resource efficiency. The data management of the network includes control of when and where data will be consumed to comply with privacy and other legal constraints. The network therefore provides flexible data observability adaptable to the requirements of the data consumer.

The distributed AI-powered execution environments will be capable of handling any AI model, including generative models. These AI models can be distributed across different locations, training collaboratively to protect data privacy and utilize resources more efficiently. In 6G, we aim for telecom AI-native solutions including the air interface, context-aware sensing and positioning. Today’s embedded AI functions in CNs and RANs will be reinforced with generative AI capabilities.

Standardization efforts for distributed AI and data in networks are underway in the 3GPP (3rd Generation Partnership Project), while intent-capable network interface standardization is progressing in TM Forum, O-RAN and the
3GPP. Ericsson is integrating trustworthy AI principles into the development process by following the EU guidelines for trustworthy AI. Overall, these developments contribute to an AI-native architecture that enables innovative network functions and services through distributed AI and data, transforming the network into a platform for innovation.

TREND #3: Transitioning toward a network architecture with horizontal layers to enable innovation for enhanced and expanded capabilities

For the time frame of 2028, our development organization is preparing for the commercialization of evolved 5G features and capabilities including intelligent RAN automation, RAN features that enhance power efficiency, and time-critical communication that enhances applications such as cloud gaming, industrial automation, the Internet of Things, extended reality and emergency services. Further, the usage of 5G access technologies is expected to extend into nonterrestrial networks (NTNs) and digital airspace during this time frame, starting with the first demonstrations of 5G enabling communication services between NTNs and smartphones. Many of these evolved 5G features will serve as templates for the fundamental 6G design choices.

Perhaps most importantly, during this time frame the network implementation will be restructured into horizontal layers. Each layer will be developed and specified according to its own pace and merits, while retaining the possibility to scale up for global adoption. This includes service-based CN and cloud RAN solutions both fully utilizing cloud-native and AI technologies. The architecture also includes open interfaces such as the air interface (New Radio) and the fronthaul interface (Next Generation – Lower Layer Split). The open network interfaces provide an opportunity for network modules to be developed separately, while at the same time being interoperable with other modules in the network. The horizontal architecture provides a more flexible, adaptable and customizable network to address the broad range of needs and requirements.

Entering the next decade, the 6G network will bring both enhanced and expanded capabilities. Built on evolved 5G, it will have unique capabilities including wider spectrum range, extended coverage, faster data speeds and lower latency for increased network capacity. Examples of future 6G technologies include full duplex radio solutions and larger massive MIMO arrays. These solutions also enable reuse of the current site grid for the new centimeter-spectrum band as well as coverage and capcity improvements in the uplink. Further, the 6G network will provide enhanced mobility with increased relative speed between devices and radio base stations, both on the ground and in the air. Crucially, 6G will be built according to a lean design approach that optimizes energy efficiency by integrating zero-energy devices and delivering fully dynamic power management that enables scalable power consumption adapted to the network traffic.

Expanded capabilities encompass intregrated radiobased sensing utilizing radio transmission to detect radiowave propagation in the physical environment. Sensing is enabled by the reuse of higher frequency and wider spectrum bandwidth as well as the infrastructure of the 6G network. The resulting sensing data includes accurate three-dimensional and real-time information about the physical world to enable applications such as digital twins and spatial maps. Integrated sensing and communication (ISAC) allows for synergies between communication and sensing functions, where the data obtained from sensing can enhance communication (through adaptive beamforming based on environment mapping, for example) and vice versa.

The 6G network will be capable of handling dynamic and temporary deployment needs that address the versatility of new use cases or specific needs, making it possible to support flexible levels of resilience and reliability, for example. Another example of this is the ability to meet specific deployment requirements such as on-premises processing, interworking with other systems and the integration of temporal or mobile nodes.

Finally, trustworthiness, resilience and availability are all essential characteristics of a secure, high-performing network as well. At Ericsson, we are using zero-trust principles to develop a zero-trust architecture (ZTA) that secures network resources and data across all elements, including identity, credentials and access management, as well as end points and their interconnecting infrastructure. The ZTA integrates security components, workflow orchestration and access policies that strictly adhere to these principles with the objective to protect all data and resources, regardless of whether they are located on premises or in the cloud.

TREND #4: Pursuing innovations over the next decade that enhance performance, sustainability and security

Novel technologies that make it possible to achieve both high performance and energy efficiency, including the diversification and specialization in hardware, are an important research priority since mainstream processors can no longer be relied upon to deliver the performance improvements that Moore’s law dictates. New processor architectures, such as neuromorphic and quantum processors, are expected to be available for commercial use for specific use cases within the next 10 years. Further, integrated photonic interconnects will increase data rates, significantly reduce energy consumption and enhance flexibility in resource utilization to meet the growing demand for high-speed data communication in future networks.

The slowing down of Moore’s law will also drive requirements on more performant real-time systems-level programming. To innovate, software developers and system architects will need to develop lower-layer software and will therefore become more adept at exploiting the capabilities in the new processor and memory architectures. Besides real-time performance, increased focus will be put on capabilities such as security and availability in hardwarenear software. One example is low-level programming languages such as Rust, which offer both memory safety and fine-grained control over system resources: two features that are increasingly important for writing performant code without sacrificing safety. In the case of AI, the focus will be on prioritizing software and algorithmic innovation, utilizing specialized hardware, and the efficient use of available computational resources.

Conclusion

Ericsson is committed to building high-performing differentiated networks that stimulate growth and innovation by exposing network assets through application programming interfaces that are accessible to developer platforms. This approach offers the advantage of catering to a diverse range of application needs, from best-effort to guaranteed connectivity, and supporting performance-based business models to deliver optimal network performance at the right cost.

The combination of 6G and AI is fundamental to this journey. I often explain it this way: the 6G access network is the muscle with unlimited reach that will power a vast array of applications and services, providing differentiated, robust and high-speed connectivity. Meanwhile, the 6G embedded network AI is the distributed brain of the system, bringing adaptability and efficiency to 6G, controlling and optimizing its full operation.

As we lead the way into the next era, Ericsson aims to spur innovations that enhance the lives of people around the world, make businesses more prosperous and foster long-term, sustainable economic growth. My colleagues and I could not be more excited about the plethora of opportunities that lie ahead. Whatever your role in the ecosystem, we welcome you to join us on the journey.

Further reading

• Ericsson Technology Review, Beyond bit-pipes – new opportunities on the 6G platform
• Ericsson Technology Review, Service quality monitoring – an essential tool in the digital economy
• Ericsson Technology Review, Evolving service management toward intent-driven autonomous networks
• Ericsson white paper, Co-creating a cyber physical world
• Ericsson white paper, Zero Trust Architecture for advancing mobile network security operations
• Ericsson Technology Review, Networking trends 2023: Building the platform for next-level digitalization

About the author

As Group CTO, Erik Ekudden is responsible for Group Strategy and Technology. His extensive experience of working with technology leadership globally influences the company’s strategic decisions and its investments in 5G, 6G, edge computing, AI, augmented/virtual reality and the Internet of Things. Ekudden’s leadership builds on his decades-long career in technology strategies and industry activities. He joined Ericsson in 1993 and has held various management positions in the company, including Head of Technology Strategy, Chief Technology Officer Americas in Santa Clara (US), and Head of Standardization and Industry. He is also a member of the Royal Swedish Academy of Engineering Sciences and the publisher of Ericsson Technology Review.