5G mobile networks: A driver for edge computing

Recently, a striking report published by Omdia and Canonical highlighted that 86% of communication service providers (CSPs) are optimistic about the future of edge computing on telco networks. 

This is a market that is expected to grow substantially in the coming years, with our report shedding light on the motivation that CPSs are drawing from the enhancements that 5G networks will bring. However, the report also highlights that standalone 5G networks have not been widely deployed yet and attributes this fact as a major reason why only limited 5G edge computing deployments exist today.

In this blog, we explain why edge computing matters and is relevant to CSPs, and how 5G networks will drive the rise of edge computing in CSP infrastructure. As 5G edge computing is also highly-relevant to private mobile network (PMN) use cases, we then also elaborate on how PMN deployments will benefit from wider adoption of stand-alone 5G technologies. Finally, we will introduce Charmed Aether SD-Core, an automated open source 5G core network for PMN edge computing use cases.

Edge computing for telco

Edge computing is an approach in which CSPs host computing applications at the edge of a wide-area network. This brings them closer to end users and the devices that consume the data delivered by those applications. This brings significant improvements in data delivery performance and quality of experience.  

In a nutshell, with edge computing, the distance between a user application and the end user is significantly reduced. This is because information doesn’t have to traverse a long route between a central cloud and the edge locations where users connect to the network. Serving applications directly at edge clouds enables much quicker interactions between user devices and the applications hosting data services. This is why edge computing is popular for many time-sensitive business applications, and increasingly becoming indispensable for mission-critical ones. 

A benefit on top of speed is the relevance of collected information. Some business applications only require local data, such as location-based services, as opposed to an aggregate bulk collected from multiple locations over a wide geographical area. Having highly relevant and local data reduces pre-processing operations which would otherwise be needed to prune collected information centrally and then store and serve processed data remotely for each and every edge site, which is significantly inefficient. 

All in all, edge computing is certainly a technology that customers of mobile services greatly benefit from, but this will be truly realised only when CSPs scale up their deployment of edge clouds. Currently, more than 60% of respondents report having less than 30 nodes available for edge applications – this is far from sufficient if edge computing is to truly take off.

Why is 5G important for edge computing?

Operators continuously seek cost-reduction techniques, which they hope to achieve with a transition to cloud-native infrastructure. They also look for areas to invest in to generate more revenue growth, and as the report points out edge computing presents an opportunity: CSPs want to host more and more applications at the edge, and provide many more business services of different types with premium connectivity. This is where 5G becomes both a driver and an enabler: it provides the means to deliver the performance levels required by different types of business services, and motivates operators to invest in edge computing. 

5G is a technology enabler, as it brings a very flexible and cost-efficient way to deploy and run a mobile network as a collection of cloud-native network functions running on common-off-the-shelf (COTS) hardware hosted at data centres. It has a highly modular architecture based on micro-services and service-based interfaces, allowing for software-driven deployment patterns across a multi-cloud environment. 

By disaggregating 5G network functions across an operator’s network clouds, it is possible to run a mobile network in many innovative ways. The functions that process user data can be dynamically provisioned at edge sites and scaled as needed to meet varying levels of traffic volumes. They can also be replicated and grouped to form network slices, which are offered as dedicated portions of a mobile network. Each of these slices can then be individually configured according to specific traffic characteristics of a data service and its quality of service requirements. 

In summary, 5G is a major driver for edge computing. It offers a much more versatile deployment architecture with components that can be pushed to edge clouds to efficiently deliver data to local business applications with various traffic characteristics, across a set of key performance metrics. 

Our telco report shows 5G will drive edge computing

As with any new technology, it takes time before operators will fully deploy 5G mobile networks. Whilst we have some of them up and running today, the Omdia report highlights that true 5G potential hasn’t been achieved yet. This is attributed to the fact that most of these deployments are based on an amalgamation of 4G and 5G components, which is what the industry calls non-standalone (NSA) 5G. 

This has inhibited the development of edge computing, because a 4G or a 5G NSA network cannot deliver the same level of performance a standalone (SA) 5G network built with purely 5G components can. With every generation of a mobile network, we get better capabilities to deliver better performance, i.e. higher data rates and lower latency.

The report notes that CSPs are already progressing with their edge computing deployments targeting completion within the next couple of years. It also highlights the significance of what 5G SA will bring to edge computing in terms of better communications quality and more use cases enabled. This makes 5G a prime enabler for edge computing, making it even more widespread, leading to many more edge computing sites being deployed at CSP networks. This is an outcome that CSPs strive to achieve, as they heavily invest in cloud-native technologies to run their 5G networks and also host edge computing applications. The report accurately captures the optimism that CSPs have for edge computing.

CSPs have chosen open source in edge computing, according to the latest telco report

The report highlights that 80% of CSPs today are using open source technologies, with the main use case being in edge infrastructure. This is because CSPs understand that using open source software brings significant cost-efficiencies in deploying, running and maintaining their infrastructure. 

Consider a wide area deployment with thousands of cell sites. With edge computing capabilities delivered to all these sites, hosted across hundreds of edge clouds, a CSP will need to scale and deploy resources cost-effectively. With secure and trusted open source solutions already available, this is a great opportunity for CSPs to finally realise what 5G can provide them with: a disaggregated and highly distributed network with integrated edge computing clouds, serving their business applications at scale across a wide area.

Edge computing for private mobile networks

It is not surprising that private mobile networks (PMN) have a lot in common with edge computing. A private mobile network is a full mobile network deployment that is owned by an enterprise, deployed across their business sites to provide premium mobile network connectivity for their business processes. It leverages edge computing technologies heavily, since business services run infrastructure that is adjacent to the deployed PMN infrastructure, hence acting as edge service for the enterprise. 

Both PMNs and edge computing technologies require an efficient infrastructure to host various types of business services, and both will therefore benefit from what 5G can provide on a cloud-native infrastructure. The report draws a parallel between edge computing and PMNs, and suggests that PMNs are a great market for edge computing to further enhance user experience. The report highlights that enterprises that deploy a private mobile network are keen on combining it with edge computing as one of the most requested add-ons to private networks. In fact, 34% of the survey respondents say that they also acquired an edge computing solution besides their private mobile network. 

A 5G PMN, as a dedicated mobile network for an enterprise, can efficiently host business applications for a wide range of industries, including manufacturing, mining, retail, hospitality, and many more. With higher data rates, much more reliable connectivity, and support for various application types, 5G is the key for a growing PMN market, and its growth pattern will parallel that of edge computing. This is because both solutions share the same underlying goal: to realise the benefits brought by cloud-native computing and meet the demand for new and innovative network services, whether it is for the consumption of end customers or for the benefit of an enterprise’s own operations, and the need for high quality of experience. This is not surprising, as all the aforementioned industries run computing applications which now effectively become edge computing services next to a private mobile network deployed on-site, directly benefiting applications with stringent performance requirements.

Charmed Aether SD-Core: Open source 5G for enterprises

Aether is a Linux Foundation project aimed at delivering open source private mobile networking software for everyone. The project first started at Open Networking Foundation (ONF), the pioneers of software-defined networking. Now part of the Linux Foundation, Aether has a broad reach in the open source community. Aether’s SD-Core sub-project distributes a software-defined open source 5G core network implementation, which is a perfect fit for any enterprise that aims to deploy and run their own 5G network. 

With Canonical’s Juju software orchestration engine, deploying and operating complex software, like Aether, is no longer a challenge. Juju provides automation in software lifecycle management with enterprise-grade software operators, called charms. Charmed Aether SD-Core is a set of charms that anyone can use to easily deploy Aether’s SD-Core 5G core network on Kubernetes as a cloud-native mobile core. 

Charmed Aether SD-Core is a perfect fit for enterprises that want to run their own 5G network, but lack the skill set and know-how on how a mobile network operates. Thanks to Canonical’s Juju engine, all the complexity is handled by the charms, leaving system administrators with a set of simple commands. Users also get a graphical user interface where they can add their devices to the network as they wish.

Juju comes with control and user plane separation (CUPS) capabilities. This means that the control plane, which is the brain of a mobile network, can be separately deployed, either on-site or on a public cloud, as needed and preferred. The user plane function (UPF), which is the network function that deals with service data traffic, can be flexibly replicated and scaled as needed at any other location(s) on demand. This is what Charmed Aether SD-Core brings to enterprises: flexible and easy deployment and management of their fleet of 5G UPFs across their many sites, right next to their edge computing applications.

Summary

5G does not only bring a new generation of mobile networking software, but also a paradigm shift in how we deploy and run mobile networks. This shift enables flexible deployment of data processing components, bringing them closer to the edge of the network, and enables much faster data delivery to cloud-native edge computing applications that can now be deployed at edge clouds at scale. Adoption of edge computing and leveraging its true potential will depend on and is driven by widespread deployment of 5G networks.

As distributed and disaggregated networks, 5G requires a suitable cloud-native infrastructure to run on. With many more edge cloud sites now required to host edge computing applications, cost-effective deployments are essential to get higher ROI for CSPs from their edge infrastructure. Open source provides the means to achieve cost-effective deployments with carrier-grade software, and CSPs recognise this. 

Private mobile networks based on 5G technologies will also greatly benefit from the enhanced data rates, reduced latency, and support for a variety of traffic types, all brought by 5G. Canonical’s Charmed Aether SD-Core provides an open source 5G core network with full automation that runs on cloud-native infrastructure, providing enterprises with a cost-effective and simple way to run their own private mobile network on COTS hardware, boosting the efficiency of their business services.

Contact us

Charmed Aether SD-Core is fully open source. If you want to try it out, check out our documentation pages today and get in touch about your private mobile networking needs.