Cedric Gegout
on 30 September 2024
Cédric Gégout (VP of Product at Canonical) reflects on a recent research report that Canonical published in collaboration with Omdia. The report uncovered that despite the lack of widespread 5G standalone infrastructure, enterprises, industrial actors and service providers are preparing for 5G edge computing to produce returns within two years. Our report reveals that they are optimistic about using virtualization technologies to capitalize on the demand for 5G edge computing, even in the absence of 5G SA. We examine the drivers behind this optimism and the steps that enterprises are taking to capture the growing demand for 5G edge computing.
The mood has shifted
As a product-oriented individual, I see business value as part of any technical solution to a problem. I often think about how setbacks affect our approach to meeting customer expectations. In the context of 5G, it’s tempting to assume that enterprises are frustrated by not having full-featured 5G networks at their disposal to boost their businesses. After all, since the beginning of 5G implementation in 2020, most deployments have been non-standalone, meaning that they continue to rely on the 4G core as a foundation.
By contrast, as of 2023, just 19% of 5G deployments are standalone. 5G Standalone Core (5G SA) refers to a complete deployment of both the 5G radio access network and the 5G core network, unlocking full 5G capabilities. It supports advanced features for tailored virtual networks, IoT fleet management, and low-latency application support. Until now, its absence has been holding back strong progress on 5G edge computing.
Despite the lag in 5G SA rollout, 96% of communications service providers (CSPs) report that they expect to launch fully commercial edge computing deployments within 2 years. What’s inspiring is that businesses themselves are now taking the lead in driving 5G deployment for their own growth, by making use of existing network resources in the absence of 5G Standalone.
Enterprises are forging a new path in the face of customer demand
So what’s behind this optimism? The appetite to wait for the underlying infrastructure to be updated has been replaced by an appetite to completely rethink how we use existing infrastructure, through virtualization and containerization, to deliver the benefits of 5G computing at the edge. This decision isn’t driven by frustration – in fact, the contrary is true. Customers want 5G at the edge now, and CSPs are already reporting revenue generation from premium connectivity features and end user applications, two services that deliver expanded coverage and remote management to edge devices.
Use cases as diverse as remote monitoring systems, drones, and autonomous vehicles are already being rolled out to consumers on the current network. The opportunity to capture this market is open to anyone who can offer a higher performance network, or rely on it to deliver their businesses. For instance, as CSPs see 5G edge computing as a solution to meet growing customer needs, many 5G infrastructure vendors are working to deliver turnkey solutions that address these technical demands.
I think this is a great example of the “tug of war” that exists between demand generation and demand capture. CSPs and enterprises expected 5G standalone to generate demand for 5G edge computing, but the demand is already out there and is about to generate a shift in the network rollouts.
A paradigm shift in radio access networks
What does this significant change in business thinking and the underlying infrastructure assumptions mean in practice? It means diversifying your options to create an infrastructure that can operate at the edge without 5G Standalone (SA), but which remains standing and ready to harness it when it arrives. In fact, 51% of respondents cited 5G SA as a key driver in the growth of 5G at the edge, despite the slow rollout – so it’s clear that whilst enterprises are keen to move forward before 5G SA is mature, that does not equate to giving up on the benefits of the technology.
From the 5G architecture, network disaggregation is becoming the key step for advancing 5G edge capabilities. This refers to the separation of network functions into distinct components and the decoupling of hardware and software functions. The goal is to move away from a fixed, rigid architecture with inflexible roles and responsibilities (usually managed by one vendor), to an agile architecture that is composed of interchangeable components that can carry out their functions independently and potentially by multiple vendors. Converting a traditional radio access network (RAN) into one that runs with virtual network functions enables organizations to benefit from cloud native infrastructures that meet the requirements of the modern applications and distributed systems.
Towards composable, interoperable and flexible networks
In practical terms, this means that all users of 5G edge computing are freed from vendor lock-in and can draw upon cloud native deployment practices. They can deploy network functions from different vendors and replace components without overhauling the system. This also frees them to use open source components instead of proprietary closed-sourced ones. In addition to the licensing cost savings this entails, by nature these components are composable, interoperable and customizable. This modularity offers enterprises the freedom to quickly and effortlessly distribute their resources across networks in line with shifting demands.
We’re not referring to market forces here, but to practical technical dynamics, and real-time shifts in network usage by end-users. Service platforms can scale with demand at the edge by deploying capabilities closer to the end-users, as and when needed. This ensures that high standards of service are maintained and that no resource is wasted. In essence, demand capture in real-time.
Diversification meets unification
By experience, I know that interoperability is one of the toughest challenges. A set of components that work independently of one another must, simultaneously, seamlessly work together while delivering cost-effective performance. Indeed, 44% of our respondents cited this challenge as a barrier to open vRAN adoption.
As edge computing becomes more widespread, deploying network functions across different clouds on different nodes will become a challenge. Establishing a common denominator cloud for your components to rest upon eliminates the need for multiple stacks and offers a vendor agnostic layer of modularity for your network. This is a great deal simpler than retaining siloed, proprietary stacks and manually establishing integrations between them.
Open source cloud platforms have an important role to play here. You don’t want to escape from vendor lock-in at the network level, to then reimpose it at the cloud level. More importantly, when 5G standalone core becomes more widespread, it will require a cloud-native approach that supports higher speed and capacity as well as more complex features such as network slicing and data processing live migration (the essence of edge computing). Open source software has interoperability baked in from the outset, shifting the focus for enterprises away from building a stack and towards assembling the cloud layer they need from existing, best in class open source components.
The rise of containerization
As a technology enthusiast, I’ll be the first to admit that we can often lose sight of the bigger picture and focus on the technical benefits – which is why it’s good to examine how 5G edge computing intersects with wider industry trends. The story of 5G edge computing is part of a wider story about maximizing the way we use our resources. This is the same story that has given us first virtual machines, and then containers.
Our report shows that currently, virtualized network functions (VNFs) and containerized network functions (CNFs) coexist, but that CNFs will become predominant by 2029. Enterprises will need to continue using platforms that support both VNFs and CNFs in the medium-term. However, as 5G edge computing matures, we can expect the retirement of VNFs as a means of freeing up resources and increasing efficiency. Indeed, CNFs are quicker to boot, lighter in terms of resource utilization, and easier to scale. This makes them a logical choice for those wishing to create an agile network that can adapt to user demand. However, the investments required to refactor the VNFs into CNFs are still high without immediate return.
A rising tide lifts all boats
As mentioned, edge computing is an essential part of how edge devices work. It follows that increasing the coverage and performance of the network offers remote enterprises more stability. It also gives them greater freedom to innovate outside the confines of the existing networks.
Most enterprises see an opportunity that goes beyond appealing to edge users. The enhancements required for high performance 5G at the edge will improve the way the network functions as a whole. This is akin to the metaphor of a “rising tide that lifts all boats”. The benefits of lower latency, improved connectivity and energy efficiency must be achieved across the entire network (public or private), including within the core, in order for them to reach the edge. The paradigm shift offered by the transition to an open and virtual RAN offers a chance for a sweeping modernization of the network. In turn, this offers us the chance to deliver a new, elevated standard of customer experience.
And that’s something which will not only enable enterprises, service providers and industrial actors to capture today’s demand, but to build a next-generation network that generates demand with tomorrow’s innovators.
Access the full report and get closer to the edge
In the full report, produced in partnership with Omdia, we get close to the barriers and motivations driving the growth of 5G edge computing. We venture beyond the content of this blog and cover topics around automation, orchestration, distributed computing and private 5G networks.
