- After rolling out automation on its own fiber network, Google Cloud is looking to help telcos do the same
- Its new Autonomous Network Operations framework details how telcos can use its products and leverage partners to build the foundations for autonomous operation
- More on this front is in the works, with AI agents for specific use cases on the way
Google Cloud executives told Fierce earlier this year that the company was well on its way to fully autonomous operation of its sprawling fiber backbone network. Now, it’s looking to help telcos head in the same direction, this week handing them a few pages straight out of its playbook.
The tech giant just published its Autonomous Network Operations framework, which provides telcos with a blueprint detailing what AI, infrastructure and analytics products (from Google Cloud, of course) they can use to whip their networks into shape for autonomous operations. The framework also comes with a list of ecosystem partners (including Amdocs, Ericsson and Nokia) telcos can turn to for help.
“This isn’t just about buying software. It’s about adopting a new operational model,” Google Cloud Product Manager Naresh Rao said in an explainer video.
Play-by-play
The framework centers around three core principles: data as the foundation, proactive intelligence based on said data, and declarative operations based on data and Kubernetes control loops. And of course, Google Cloud notes it has plenty of tools – like Spanner, Spanner Graph, Big Query, Big Query Graph and more – to help build these pillars.
So, where do operators start? Angelo Lubertucci, Google Cloud’s head of Industry for Telecom, told Fierce the first step is unifying data and creating a pipeline to get the right data from on-prem to the cloud, all with the goal of building a single source of truth that can in turn be used to create a digital twin of an operator’s network.
And while it may be tedious and time consuming, this isn’t a step they can skip.
“You need a digital twin to become truly autonomous. You’ll never get to Level 5 autonomy without a digital twin,” he told Fierce.
Why? Because, Libertucci said, you need that replica of the network to pinpoint the root cause of issues, test AI agent recommendations without blowing up the network and so on.
The biggest challenge telcos face in solving the data equation is looking at what’s available to them and deciding on what data really matters for the customer experience and then what AI model to apply to that data.
“That’s where our opinionated view comes in handy,” Libertucci stated, noting sometimes a prescriptive approach can actually help.
Expert advice
But why should telcos trust a cloud provider when it comes to network advice?
Well, Libertucci said Google Cloud ate its own dog food when building its fiber backbone, building AI and automation in from the start. Thanks to this, the company operates its network with a fraction of the staff a traditional telco employs, he added.
Plus, on the brownfield side, it’s already been working the framework with the likes of Bell Canada, Deutsche Telekom and Telstra.
Bell Canada, for instance, was able to reduce the number of service calls coming in for RAN anomalies by 25% after teaming with Google Cloud and using historical data to build an algorithm capable of spotting issues.
Libertucci added that the framework it debuted on Thursday is just Phase 1 of the playbook. Google Cloud, he said, will “continue to develop agents for specific use cases” and so on, going forward.
Will this give Google Cloud a leg up in its battle with AWS for telco cloud workloads? We'll be watching closely.