Scale on a different scale
The telecoms sector is already moving to cloud-native infrastructure. Designed to accommodate explosive growth in IoT devices, it promises a giant leap forward for consumer and business customers – both in terms of the connection speeds they expect and the number of networked devices and applications they will bring online. To make things even more challenging, 5G also specifies higher performance requirements for the latency and throughput of the core network.
What this all means for providers is an exponential increase in the amount of network traffic, alongside a drop in the latency with which compute services must be delivered – which entails serving and caching more applications at the edge of the network. In short, 5G will give new meaning to the concept of delivering at scale.
451 research explains: The concurrent process of global aggregate scale and rollout timing of 5G make telecom operators the cloud-native ecosystem's most important revenue generators and discerning customers. The unique characteristics of telecom (such as stateful applications, IPV6 support, SLAs) will require a degree of specialization for suppliers, and the need for skills augmentation, developer enablement and knowledge transfer will open ample service revenue paths.
Why containers matter
To deliver against these requirements, the industry needs a cost-effective infrastructure platform that can deliver vast amounts of compute resources at the edge. But that’s not all.
With all this increased scale comes an increase in the complexity of the network itself. We have already seen a shift from networking hardware to Network Function Virtualization (NFV), resulting in software defined networks that can be reconfigured more easily. Now the industry is moving beyond virtualization and its associated costs, to Network Function Containerization (NFC). That involves running the network control plane in containers – and for today’s telco industry, containers means Kubernetes.
Kubernetes is perfect for 5G
Kubernetes is the natural solution for handling the containerization of core network functions, as well as for shifting thousands of applications and services to the edge. Now proven in enterprise environments across the globe, Kubernetes is today’s standard container platform. It is perfect for the kind of stateless management workloads 5G requires. And because it automates so much of the operational plumbing, it can cut costs and complexity too.
“Our application teams have more or less all the basics in their hands to start deploying and running their application in our very specific environment. We are running everything on-prem. We run the workloads also traditionally in many locations. We have a handful of what we call core locations, 20 to 25, then there are near edge locations and there are more than 10,000 edge locations.” Vuk Gojnic, Deutsche Telekom
Listen to the full podcast episode “Kubernetes at Deutsche Telekom - GitOps at the Edge”
“Traditionally, virtualizing multiple Kubernetes clusters has relied on heavyweight and expensive virtualization approaches. Instead, Weaveworks has demonstrated an approach based on KVM and Firecracker (known as microVMs), which provide lightweight virtualization capabilities and significant benefits for running multiple virtualized Kubernetes clusters in many environments, including edge.” Richard Case, Tech Lead on Bare Metal Kubernetes, Weaveworks
Read how Weaveworks, in collaboration with Deutsche Telekom and AWS, significantly improves the efficiency of deploying Kubernetes across bare-metal pools. Liquid Metal: Multi-Cluster Kubernetes on Bare Metal with microVMs
The software is just the start
Recalibrating around the requirements of 5G means looking at process as well as infrastructure. It means finding a way to reduce the labour overhead involved in running tens of thousands of clusters at distributed locations. Processes and policies must be standardized to ensure security, compliance and access for the right teams – not to mention compatibility with legacy environments still running elsewhere in the business. Crucially, you need to introduce the new infrastructure and processes seamlessly. Which means no learning curves. No expensive training. No disruptive trial and error. The answer is to combine the proven container infrastructure of Kubernetes with a development toolset that your app teams already use: Git.
Go further (and faster) with GitOps
Organizations will need to standardize on how they manage and operate applications in order to take full advantage of Kubernetes orchestration capabilities. GitOps can provide not only the necessary scalability and security but also the remote management that is needed in a 5G world.
GitOps is a set of tools and processes that brings together Kubernetes and Git to create a self-service IaaS platform for developers and operators. More than just another interpretation of DevOps, it is a set of technologies and best practices that offer everything you need for massively scalable edge computing.
Deutsche Telekom’s Till Stenzel recently talked to Weaveworks founder Alexis Richardson about GitOps - and why it’s such a good solution for telcos grappling with 5G.
“We see the GitOps model really as a key framework to achieve that level of automation and consistency across all of the workloads and platforms that telcos have to operate.” – Till Stenzel, Managing Director Advisory, Deutsche Telekom Capital Partners
What makes GitOps different?
- Productivity: By automating so much of the CI/CD pipeline and giving developers a self-service platform, it reduces complexity and operational overhead.
- Speed: The speed with which changes can be deployed and rollbacks made leads to higher availability overall. Learn more on the correlation between deployment speed and stability
- Standardized workflows: The entire running system - code and operational instructions - is held in version control. This provides full observability over large, distributed deployments.
- Scalability and reliability: Whole clusters can be rolled back or destroyed and brought back, in minutes. This can all be achieved at massive scale, making it perfect for the 5G use case
- Security for regulated environments: RBAC means policies and access can be controlled at a granular level, giving your app teams a secure self-service platform with all the guardrails you need. Git’s record provides full audit trails for compliance. Learn more on how GitOps improves the security of your deployment pipeline.
Putting it into practice with Weave GitOps
GitOps was pioneered by Weaveworks. Backed by leading telcos like Deutsche Telekom, Ericsson, Orange and hyperscalers like Amazon, we offer a commercially supported GitOps package called Weave GitOps. Available for all hyperscalers, on-premise installations and hybrid architecture, it makes the process of adopting GitOps seamless and straightforward.
Weave GitOps is a modular platform for applications that use Kubernetes. It includes full multi-cloud Kubernetes management and support options as well as enterprise GitOps features for accelerated and progressive delivery, add-ons management, policy and fleets.
It uses GitOps to automate deployment pipelines that manage applications plus the add-ons needed to run as a ‘stack’ on a fleet of Kubernetes clusters, a perfect match for the rapid growth and massive scale experienced by Telcos.
Key features for managing Kubernetes at the edge:
- Team Workspaces which simplifies application portability and collaboration
- RBAC for policy enforcement at scale
- CAPEI (Cluster API for Existing Infrastructure) uses an innovative light-weight approach that does not require a central management cluster, making it suitable for ephemeral or intermittently connected environments, such as Edge and Far Edge.
Professional Services for GitOps
Our expertise can accelerate your cloud native journeySchedule a call