How telcos are building carrier-grade infrastructure using open source

Telco cloud implementation with Canonical and HPE

Service providers need cloud infrastructure everywhere, from modern 5G and 6G network functions running in the network core to sophisticated AI/ML jobs running on the edge. Given the sensitivity of those workloads to any interruptions, outages or performance degradations, the cloud infrastructure used by telecommunication companies needs to be fast, robust and ultra stable.

To satisfy these requirements, telcos are increasingly turning towards open source infrastructure solutions – most notably OpenStack and Kubernetes – which deliver greater levels of flexibility, improved economics and easier access to innovation compared with traditional, proprietary software.

Today’s blog examines the drivers behind this open source infrastructure trend, and explores the advantages that telcos can gain by utilising a carrier-grade open infrastructure stack alongside purpose-built, optimised hardware.

Watch the Telco cloud implementation webinar

Why telcos are moving towards open source

Keeping up with innovation is one of the greatest challenges facing telcos today. The telecommunications sector and technology landscape are continuously and rapidly evolving, and in such a competitive industry, only the most agile businesses can hope to stay ahead. Tier 1 telco operators must ensure that they remain attractive to not only external and internal customers, but also to employees, or else risk losing critical talent.

The key to staying at the leading edge is flexibility. Telcos need the freedom to seamlessly adopt and pivot to the latest technologies. But this kind of agility is antithetical to traditional, proprietary infrastructure strategies that are highly prone to vendor lock-in. 

Energy costs can represent up to 40% of a telco’s OPEX, so the ability to dynamically scale cloud resource consumption is critical to data centre economics. At the same time, telcos need the flexibility to choose the ideal hardware infrastructure for their use cases, striking the right balance between cost and performance according to their specific requirements.

These challenges, and many more, can be solved using open source cloud infrastructure solutions. Open source software prevents the risk of vendor lock-in, and broad interoperability and compatibility make it significantly easier to adopt new solutions and integrate with a much wider range of hardware, including commercial off-the-shelf products. Innovation typically reaches the open source landscape well before it is adopted by proprietary vendors, and leading open source software can match or exceed the security and compliance capabilities of traditional solutions at a fraction of the cost.

OpenStack and Kubernetes

The core technologies behind an effective open source infrastructure strategy are OpenStack and Kubernetes. OpenStack is the most widely deployed open source cloud platform in the world, offering extensive hardware compatibility for cost-effective scaling, and a high level of elasticity to support on-demand resource provisioning. OpenStack enables telcos to optimise cloud costs and build infrastructure competitive with hyperscalers, while maintaining the control and sovereignty advantages of a private cloud. 

Kubernetes is the leading open source container orchestration tool. Containers are invaluable for driving rapid, repeatable cloud deployments with reusable building blocks, and Kubernetes is the solution that makes large-scale container strategies viable by simplifying and automating container lifecycle management.

Canonical’s open source infrastructure stack

There are many different distributions of OpenStack and Kubernetes to choose from, but not all are suitable for enterprise environments. Canonical OpenStack and Canonical Kubernetes are carrier-grade solutions delivered and maintained by the company behind Ubuntu.

What sets Canonical infrastructure apart from alternative options is the total, bottom-up automation made possible by MAAS and Juju. MAAS is Canonical’s tool for automating the provisioning, management and maintenance of bare metal infrastructure, while Juju is an open source orchestration engine for automating the full lifecycle of infrastructure components and applications. 

The automation enabled by MAAS and Juju drives down operational costs, making a Canonical OpenStack cloud one of the best performing infrastructure options at a uniquely low TCO.

Additionally, Canonical’s infrastructure provides the reliability and stability necessary to prevent downtime or performance degradation for critical telco workloads. Just like Ubuntu itself, Canonical OpenStack follows a predictable release cadence with a new long-term support (LTS) version issued every two years, each of which receives 10 years of security maintenance from Canonical. And service providers looking to take advantage of the latest OpenStack features can benefit from seamless, fully automated version upgrades. 

Last but not least, both Canonical OpenStack and Canonical Kubernetes provide full support for Enhanced Platform Awareness (EPA) features, helping telcos ensure the highest possible performance for network services running in the cloud.

Implementing and operating open source infrastructure demands specialised engineering expertise that organisations may not currently possess in house. To bridge this knowledge gap, Canonical offers comprehensive design and delivery services and an enterprise-grade support package that covers the entire open infrastructure stack – including not just OpenStack and Kubernetes, but also Ceph, LXD and Ubuntu itself – under a single, cost-effective subscription. Alternatively, customers can offload operational responsibilities entirely and instead focus on their core business with Canonical’s fully managed infrastructure service.

Telco Validated Designs with Canonical OpenStack and Kubernetes on HPE hardware

Even the best open source infrastructure software cannot deliver its full potential without the right hardware. To ensure interoperability and optimal performance, Canonical works directly with leading hardware vendors such as Hewlett Packard Enterprise (HPE).

As one of the world’s foremost providers of enterprise hardware, HPE offers a broad portfolio of open infrastructure platforms that are optimised for telco workloads and designed to minimise the risk involved in telco cloud modernisation. 

HPE’s Telco Blueprints and Validated Designs are custom hardware solutions tailored to the unique requirements of telco customers. HPE conducts end-to-end validation across the full scope of components and telco-specific capabilities, and collaborates closely with partners like Canonical to pre-validate the hardware with leading software.

Together, Canonical and HPE have created several reference architectures for running Canonical OpenStack and Canonical Kubernetes on HPE server platforms. These extensively-tested frameworks empower telcos with a proven and rapid path to value for carrier-grade open source infrastructure, optimised for both cost and performance.

For a closer look at telco cloud implementation and a deep dive into real-world use cases, click on the button below and watch the joint webinar from Canonical and HPE.

Watch the Telco cloud implementation webinar