Enterprises are showing interest in the potential of containerized 5G workloads to improve flexibility, scalability and cost-efficiency. 5G workloads are designed to be highly reliable and scalable, as is cloud infrastructure. It makes sense then to deploy 5G workloads in the cloud and to take advantage of containers to add more flexibility to the equation.
It can be challenging, however, to decide which type of cloud, such as a public cloud or a telecom provider's cloud, to use for 5G functions. Also, containers aren't always the best fit for 5G workloads.
What is a 5G workload?
5G workloads -- sometimes called 5G network functions -- are the applications that help manage communications on 5G network infrastructure. They perform tasks like authenticating users and devices that connect to a 5G network and managing data that flows across the network.
Similar functions exist on earlier-generation network infrastructures, such as 4G. However, 5G workloads pose special challenges for two main reasons:
- New architecture. 5G networks rely on a new type of architecture, typically referred to as the 3rd Generation Partnership Project architecture. This architecture implements core functions and organizes them in a more complex way than preceding telecommunications networks.
- Performance requirements. 5G networks must meet substantially higher performance requirements. For example, they transmit much more data at much lower latency rates.
Telcos, or companies that operate 5G networks, must find a way to deploy 5G workloads on their networks that addresses both these challenges.
Telco clouds vs. public clouds for 5G workloads
5G workloads could theoretically run on any infrastructure if they have connectivity to the network infrastructure they help manage. In most cases, however, it makes sense to deploy workloads in a cloud rather than on premises because clouds deliver greater levels of scalability and reliability. The cloud also makes it easier to select deployment locations close to the data sources that 5G functions must manage. This proximity helps reduce latency.
While it's clear that most 5G workloads should live in the cloud, deciding which type of cloud should host them is not always as obvious. Two main options are available: telco clouds and public clouds.
A telco cloud refers to data center infrastructure that a telco owns and operates. In general, deployment in a telco cloud leads to better performance because telco clouds integrate more tightly with the network infrastructure and data flows that 5G functions help manage.
Telco clouds also offer providers more control over 5G workload configuration. For example, providers could choose to host functions on bare-metal servers, which can help boost performance compared to hosting them on VMs.
Public clouds, such as AWS, Microsoft Azure and Google Cloud Platform, can host 5G functions that connect to telco networks. The functions aren't part of those networks, which potentially lead to poorer performance. But deploying 5G workloads in the public cloud is generally faster and simpler because telcos don't have to acquire, provision and manage their own infrastructure.
There's no one-size-fits-all answer for choosing between telco clouds and public clouds for 5G. Network operators must weigh the pros and cons of each option. In some cases, they might use both cloud infrastructures simultaneously, leading to 5G workloads that are spread across a hybrid cloud infrastructure.
Pros and cons of containerized 5G
No matter which type of cloud infrastructure telcos choose, they also have an important choice to make regarding how they deploy 5G functions. For example, they can use containers or run the workloads directly on VMs or bare-metal servers.
Generally, containers are emerging as the most common option for 5G deployment because they offer several important benefits, such as the following:
- Portability. Containerized 5G workloads can move more easily between environments because containers abstract applications from the hosting environment.
- Optimized resource usage. Containers have less resource overhead compared with VMs because they don't require hypervisors or guest OSes.
- Cost-efficiency. The lower resource overhead of containers can also reduce costs because telcos have less infrastructure to pay for.
- Sophisticated orchestration. Telcos can orchestrate containers using technology like Kubernetes. This orchestration makes it easy to manage many 5G function instances at once. Orchestration technologies also exist for VMs and bare-metal apps, but most are proprietary and not as familiar to engineers as Kubernetes.
But containers raise some additional challenges for 5G workloads. The greatest is they increase complexity because they introduce tools and layers, such as container runtimes and orchestration platforms, that don't exist with other deployment models. In addition, containers can complicate security because containerized environments have more moving pieces. This leads to more potential attack vectors and increases the risk of configuration oversights that could expose resources to attack.
The bottom line on containerized 5G cloud workloads
A few things are clear about 5G workloads. The first is that most workloads benefit from running in some type of cloud. But whether a telco cloud is better than a public cloud depends on factors such as how much latency the workloads can tolerate and how much infrastructure capacity a telco cloud has.
Another consideration is that containers generally increase the flexibility and efficiency of 5G workloads. But those benefits also come at the potential cost of creating more complexity and security risks.
Most 5G workloads run in some type of cloud, and most are probably containerized. But exceptions arise because not every 5G function is the ideal candidate for containerized, cloud-based deployment.
Chris Tozzi, senior editor of content and a DevOps analyst at Fixate IO, has worked as a journalist and Linux systems administrator with particular interest in open source Agile infrastructure and networking.