GitOps across the Edge and all Clouds

[weaveworks]

Telcos are at an important phase in their journey from static 4G networks to the dynamic world of 5G that is powered by edge infrastructure. They are now faced with the reality of having to manage thousands of edge locations - far more than ever before. This complexity can slow down operations. In fact, it's already clear that telcos are facing indefinite delays in their go-to-market strategy for 5G services. The pandemic and geo-political factors aside, technological clarity in how to implement the next generation of networks is likely the biggest obstacle in the way of telcos and their 5G goals. In order to bring their services to market faster telcos need to increase their pace of operations. They need to manage complexity at every layer of infrastructure, especially at edge locations. This calls for a well-thought-out strategy that is automated, and scalable.

In this two-part series we discuss why Kubernetes is vital for telcos to set up operations at the edge, and how GitOps, with its declarative approach to infrastructure management is key to managing and scaling Kubernetes at the edge. Together, they make for a comprehensive strategy for telcos looking to rollout 5G services in the near future.

Kubernetes - Core to edge

Services at the edge will be powered by a mix of container-based infrastructure and virtualized infrastructure. Telcos need the ability to run CNFs without orphaning their existing VNF-based workloads. The solution to this is to have a common platform that can support both technologies. That's what Kubernetes is. Kubernetes can run anywhere - core to edge - making it the ideal candidate for an end-to-end infrastructure management platform. There is industry-wide consensus around Kubernetes because of it being open and vendor agnostic. Also, it has some powerful capabilities for managing distributed infrastructure. This is particularly relevant to the edge which is made up of hyper-distributed infrastructure.

The key features of Kubernetes are:

• Automated scheduling of pods on to nodes
• Self-healing pods and containers
• Mature deployment management with automated rollout and rollback
• Decoupling of infrastructure and applications
• An extensible architecture for add-on services for load balancing, security policies and more

These capabilities of Kubernetes are critical to enabling 5G. However, you still need the right approach to managing Kubernetes, especially for edge operations.

Extreme automation with GitOps

Shewta Kapur of Nokia notes that 'A highly automated CI/CD can help improve time to market by up to 70%.' She suggests that 'extreme automation' of the deployment pipeline is the only way telcos can take their services to market faster. Her view is accurate. However, to build on it, extreme automation is possible with an extension of CI/CD - the GitOps model.

"GitOps is a way to do Kubernetes cluster management and application delivery. It works by using Git as a single source of truth for declarative infrastructure and applications. With GitOps, the use of software agents can alert on any divergence between Git with what's running in a cluster, and if there's a difference, Kubernetes reconcilers automatically update or rollback the cluster depending on the case."

GitOps takes a declarative approach to provisioning container infrastructure using Git as the source of truth. It enables smaller, more frequent deployments whether that’s to the cloud or to the edge. This means faster time to market with new updates and features. Looking ahead, this battle-tested model will be used to provision CNFs and VNFs at the edge.

“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 (view the entire interview)

GitOps' many benefits

GitOps allows organizations to deploy to multiple targets simultaneously. The ability to update a service in a core cloud location and a related service running at an edge location at the same time is powerful. GitOps allows telcos to rollout precise fine-grained deployments with ease. The changes can be made by updating a manifest file in a Git repository. Once the changes are 'committed' the GitOps pipeline is triggered, and this automatically rolls out the change to the target locations.

The biggest benefit of having a pipeline to manage deployments is the control it delivers over the entire process. Typically, deployments at the edge are not as straightforward as in the cloud. A pipeline helps to factor in all the complexities of deployments at the edge.

Another key benefit that the GitOps approach brings to edge operations is fully versioned change management. With every change to the system in production being recorded as a change to a repository, it's easy to track changes and map them to exact people and resources. This is critical for compliance.  Dive deeper on how you can deploy at the speed of 5G. 

Conclusion

As telcos race to be the first to gain mainstream adoption in the 5G market, they will need to solve the major technology-related challenges ahead. In this post, we discussed two important pieces of this puzzle - Kubernetes as the central management platform from core to edge, and GitOps as the operating model on top of this platform. The benefits of GitOps are proven in the move to cloud-native infrastructure, and these can be applied just as well to edge infrastructure. Contact us for a demo of GitOps in our Weave Kubernetes Platform today.

View the full article