One of the core value propositions of telco cloud platform providers is that they offer to deliver to telcos a standardised and open, horizontal telco cloud that will enable them to deploy and operate multi-vendor, cloud-native network functions using standard orchestration and automation tools.
The key element that is presently facilitating this is the open-source Kubernetes container orchestration platform. Similar to OpenStack, there are various ‘flavours’ of Kubernetes: for example, those of Red Hat (OpenShift), VMware (Tanzu), and each of the big three hyperscalers, AWS, Azure and Google Cloud. But the difference is that, in part because of its open-source character, Kubernetes remains sufficiently generic across all its flavours to ensure that CNFs deployed to Kubernetes can be delivered and distributed across multiple clouds that use different Kubernetes versions. Accordingly, Kubernetes is a driver of multi-cloud and hybrid cloud operations: enabling NFs to be distributed, scaled and orchestrated far more easily and flexibly across multiple environments and locations – although telcos may still, initially at least, require support from SIs or platform vendors to orchestrate their services across multiple Kubernetes instances. This standardisation potentially eliminates much of the operating cost telcos have hitherto expended on private telco cloud development, and on testing, validation and integration of NFs to their private clouds. One of the vendors we spoke to even suggested that telcos should go further by using the hyperscalers as the development department for their NFs themselves. The respondent cited the example of AT&T and gave a positive assessment of AT&T’s divestment of its Network Cloud (private telco cloud) to Microsoft Azure. This included, according to our respondent, teams and activities involved in developing AT&T’s 5G Standalone (SA) core.
Network evolution through cloudification
Source: STL Partners
This could look very much like a cloud take-over of the telco network. But what counteracts this risk is telcos’ evolution of their own cloud-native operations and software practices, particularly CI/CD and DevOps. In our survey, vendors and telcos alike stressed how imperative it was for telcos to develop and manage their own CI/CD pipeline. Why is this? It is because this is the means for telcos to assert control and ownership over the development of their network capabilities and innovative services even as the NF and infrastructure building blocks they are using are becoming more generic, open and common to multiple third-party clouds, including those of other telcos and hyperscalers. Even if the development of the telco cloud and of the CNFs deployed by telcos is outsourced to platform providers, hyperscalers and vendors, the mastery of these cloud-native operational processes by telcos can ensure that telcos are not just passive recipients and consumers of the cloud products and services of others.
After all, DevOps implies an interdependence of NF software development and operation. So it is not the case that telcos will simply operate what others have developed: in the cloud-native world, operating a network successfully implies continuous development of it – continually improving and adapting network capabilities so that they can cost-effectively serve the needs of customers and applications.
For more details please see our report Telco Cloud: Short-term pain, long-term gain?