Ericsson deepened its radio access network (RAN) integration with Google Cloud, adding support for common computing platforms and tapping into artificial intelligence (AI) and machine learning (ML) to ease operator 5G deployments.

The expansion extends Ericsson’s Cloud RAN system onto Google’s Distributed Cloud (GDC) Edge platform. This was tested by implementing the Ericsson virtual distributed unit (vDU) and virtual central unit (vCU) on the GDC platform using an x86-based accelerator stack running in Ericsson’s Open Lab in Ottawa, Canada.

A vDU is software running on top of commercial hardware and located at a far-edge location near an antenna or radio unit that beams out a wireless signal. It typically handles real-time processing to manage the traffic from that radio unit.

A vCU runs between a vDU and a network core, typically in an edge data center. It acts as a nonreal-time processing unit that can control several vDUs in a network architecture.

These Ericsson Cloud RAN systems running on the GDC Edge platform can tap into Google’s Cloud Vertex AI, Big Query and other cloud services to better optimize and control their RAN infrastructure.

Manish Singh, VP for strategic alliance and key account manager for Google at Ericsson, told SDxCentral that the ability to merge the software-based infrastructure with Google’s embedded tools eased this integration work.

“We see a lot of benefit of leveraging that because we can see that the tools they’re using allows us the freedom if we want to change things around,” Singh said, adding, “we didn't have to make much changes while doing this work.”

In terms of expansion timing, Singh said the latest Google updates aligned with Ericsson’s internal sequencing “based on their readiness when it comes to using open automation, open logging, performance management, what can be performed on the suite and which we can interchange,” Singh said. “At this point it looked like Google was very receptive and ready from that perspective.”

This also plays into the ongoing evolution of broader telecom equipment virtualization efforts. Singh noted that while it would be great for platforms to begin to settle, the ecosystem is still evolving.

“We would want a steady or at least continuity or evolution versus, hey, now today it is an HP [Hewlett Packard Enterprise] with a particular thing and then a Dell with another one,” Singh said.

He did note that it’s becoming easier to deal with this fluidity due to the increasing use of containerization to package network systems and functions and the accompanying increased use of AI and ML to help with integration and deployment.

“We want to be sure that we pick that set of tools which allows this,” Singh said. “We already had it on Red Hat, for example, but now it's more about can I use the tool sets, the AI and ML, whether it's for logging, whether it's for performance management. It was a strong focus for this whole exercise that we've been going through.”

The latest integration also tapped into the Linux Foundation-hosted Nephio project. That Kubernetes-based project was initiated by Google Cloud and donated into the Linux Foundation last year.

It’s designed to provide Kubernetes-based cloud-native intent automation and automation templates to make it easier for telecom operators to deploy and manage multi-vendor cloud infrastructure and network functions across large-scale edge deployments. It sits on top of a Kubernetes substrate either directly controlled from an operator or a hyperscaler-based platform like Google Config Connector, Amazon Web Services (AWS) Controllers for Kubernetes, and Azure Service Operator.

It basically takes Kubernetes’ cloud-native orchestration expertise to allow operators to roll out and manage new services in their 5G and edge deployments. This will allow for faster onboarding of network functions into a production environment that mimics the DevOps process used by hyperscalers.

The project recently unveiled its first big release, which uses the Kubernetes Resource Model (KRM) approach to manage and deploy services instead of relying on Helm Charts that are often used to deploy applications in an enterprise context. The model for KRM package creation is based on a GitOps approach to enable a controlled approach to version management and updates.

Nephio also integrates an intent-based automation model that will make a big impact for operators.

Singh said that Ericsson was working with Google to further develop Nephio toward this goal.

“We are now looking at some of the tool sets,” Singh said of that work. “We need a product at the end of the day, so one is setting the standards, one is product. We are examining some of the tools that Google has provided to see how Nephio-compliant it is, whether it locks us in or whether it allows us that flexibility. So that's still evolving.”

Ericsson began its virtualized Cloud RAN efforts in late 2020, and released its first commercial offering in late 2021. The vendor has since been updating and validating the platform to work in different environments and conducted trials with some operators.

The latest integration also bolsters Google’s telecom-focused cloud efforts.

The hyperscaler unveiled its Distributed Cloud efforts in late 2021, and commercially released those efforts last year. It has since been validating different core and RAN systems from telecom vendors to run on the platform.

Google has touted Verizon and Bell Canada as two telecom operators using the GDC Edge platform, and more recently partnered with T-Mobile US to support that operator’s enterprise-focused 5G Advanced Network Solutions (ANS) platform.