DevOps, 5G & IoT Success
The telco industry typically sees new software going live within 1-2 years; in the world of 5G, this simply won’t cut it. It’s safe to say that the telco industry must significantly shorten its time to market in order to be considered a viable infrastructure partner for the likes of Netflix, Hulu, and other leading content/ OTT providers.
5G migrations will lead telcos to adopt DevOps principles and toolsets to expand and fully leverage software-defined networks.
The complexity and scale of these specific use cases require a shift from the current DevOps tool-chain-based approach to a more holistic approach known as Value Stream Management. Value Stream Management Platforms (VSMP) look to go beyond Agile and DevOps and focus on increasing the end-to-end flow of business value – from customer request to customer delivery.
Gartner predicts that by 2023, 70% of organizations will use VSMPs to improve flow and add value to the highly complex DevOps pipeline.
Cloudify EaaS (Environment as a Service) and its integration with Multi-Cloud, different Kubernetes platforms (OpenShift, GKE, EKS, AKS), infrastructure orchestration (Cloud Formation, Azure ARM, Terraform), and configuration management (Ansible) and CI/CD (Jenkins, CircleCI, and GitActions) can serve as a useful reference on how to turn VSMP from concept to practical reality specifically as it relates to the orchestration of network and 5G environments.
The following describes the Cloudify value stream approach to managing various 5G and edge network use cases.
Use Case 1: ‘Agile-First’ Approach to 5G Network Slicing with AWS & Cloudify
We take an agile-first approach in this example by applying DevOps practices to 5G network slicing management.
The first step taken was to use a cloud-native approach to manage the entire 5G network. In this case, the AWS team used AWS cloud services (such as EKS) both as the central site and corporate site/ edge to deliver a fully cloud-native 5G infrastructure. A Nokia cloud-native 5G stack was used to manage the 5G network on top of EKS. Management of the network slicing workflow was delivered via the AWS Pipeline.
Cloudify is used here as a service orchestration layer that integrates with all relevant AWS services to deliver end-to-end service automation to the operator. In this example, we used AWS Lambda as the generic interface that calls those underlying services.
The agile first approach to deliver 5G network slicing
Agile 5G: Check!
Months & Years Become Weeks, Days & Hours.
In this AWS project, we were able to deliver the entire use case including a continuous update in just a few weeks. We were able to iterate, add new features and fix bugs with a distributed team that was spread across India, the United States, and Israel because we all had access to the same environment. We were also able to add new features and fix bugs within just a few hours!
The integration between Cloudify and AWS Lambda provided a generic interface that allowed us to access any AWS service and API, add custom logic around the API, and even abstract the chosen services into fewer API calls. This played a key factor in achieving this degree of agility.
Use Case 2: Managing Distributed Workloads & Infrastructure Across a Cloud-Native Edge Environment
In this use case, we use EKS, AKS or GKE as the central Kubernetes controller and StarlingX as edge infrastructure. Cloudify is used as the multi-cluster Kubernetes manager that can manage workload across thousands of edge devices which are essentially Kubernetes clusters.
Cloudify also supports a mixed workload scenario between cloud-native and OpenStack VM based workloads.
This is done by utilizing the StarlingX APIs (control-plane) for discovering the various edge sub clouds and communicating directly with these subclouds to provision resources, via a placement policy that defines which workloads run and where.
You can read more information about this use case and see a live demo on StarlingX and Cloudify post.
Use Case 3: Using an Open Source-Based Stack to Manage 5G Network Slicing
In this use case, Cloudify is being used as a service orchestration that can manage network slicing between a traditional on-prem data center and a variety of edge devices that are not always cloud-native.
You can read more about this use case via this Orange Lab post: ‘An end-to-end open-source mobile network with traffic prioritization mechanism’
Here, we’re using an API gateway to map the 3GPP API calls into Cloudify.
On the southbound interface we are using the Cloudify plugin approach to call 5GICE using TMF standard as well as other legacy network services using the generic Rest and SSH plugin.
Continuous Update & Management Across Development & Production Environments
Cloudify integration with CI/CD platforms such as GitActions, CircleCI, and Jenkins allows users to trigger entire deployment and service configuration updates just by updating the service definition file – a blueprint. This update (commit) will trigger a deployment workflow into a test environment; once the tests are completed successfully it will merge into a master and into the production system.
Cloudify provides a templatized environment definition that allows setting up a development environment optimized for cost, and a production environment that is optimized for availability – consequently optimizing the cost between the development and production environment.
In all three use cases above we use DevOps best practices to deliver agile 5G and edge network management. In the first use case, we leverage AWS as the infrastructure and DevOps platform, achieving both agility and cost efficiency by taking advantage of the full infrastructure capabilities provided through the public cloud.
In the second use case, we use a Cloud Native approach to simplify workload management across many edge devices. Here we were able to provide a more ‘infrastructure agnostic’ solution, spanning across multi-cloud and edge devices via a common control plane.
In the third use case, we can see open-source stack managing devices that are not necessarily cloud-native to effectively manage 5G network slicing. We also shared how it is possible to automate continuous delivery across development and production environments in a generic way that can be applied to each of the use cases in a similar way.
The above examples demonstrate the flexibility of Cloudify as it works across the various use cases, as well as via built-in integrations with cloud infrastructure, Kubernetes, and CI/CD. With Cloudify in play, organizations can move up the stack of the value chain and focus on end-to-end use and business value including infrastructure cost optimization and managing workloads across multiple clusters and domains – they can even launch new business services and govern them accordingly.