Orange faces network automation challenges with core vendors

Orange is experiencing challenges as it works through the process of designing cloud native, automated operations with its vendors. Read this extract from TMN Q's cloud native telco interview series.

TMN’s upcoming issue of TMN Quarterly features an exclusive series of interviews looking at the topic of network automation and AI. We speak to AWS, Ericsson, Netcracker, Orange, Snowflake and TM Forum. This is an extract of our interview with Aroussia Maadi, Automation for Network department director at Orange.

In it, Maadi outlines how Orange is working with vendors to enable true cloud native operations – a process that has involved some “very tough debates”.

You can sign up to receive your copy of the magazine here.

Orange took all the work that was done by TM Forum, adapted it slightly to its needs, and set itself the “very ambitious” goal of achieving level four Autonomous Networks, as defined by the TMF AN manifesto, in 2025.

It will apply this Level Four ambition to two main processes: monitoring and change, across RAN, backhaul and core domains. It has begun gap analysis, and maturity assessments in France, Belgium and in some of its African country operations.

This is proving “challenging and even impossible” in some domains, according to Maadi. For instance, although there is some maturity in the transport domain where Orange introduced Cisco Network Services Orchestrator four or five years ago, the core network is proving very tough.

“It’s much more complicated because we are not only talking about operations, but also how we are building the functions in the core. We had some experimentation on the automation of deployment, and common pipelines, for VNFs. We are just beginning the journey for containerised network functions.”

That has involved some “very tough debates” with Nokia and Ericsson – the vendors that Orange has selected to deliver containerised cloud core functions.

“With Nokia and Ericsson it is quite complicated because there’s a mindset shift also on their side. You know Nokia is invested in Sylva, in Nephio’s Kubernetes based approach (Footnote 1). But when we come to reality – so we are now deploying functions for Orange France, for example – they are delivering operators packaged like before.

“So they put the network functions in containers and deliver them, and they do not adapt by design to have independent operators, Kubernetes operators, where we have a separation between the telco parameters and the system parameter.”

Pikeo was easier because we had a more cloud native core. We didn’t have Nokia and Ericsson. Now we have real life and it’s more complicated.

The move to automate a production-ready cloud native core comes after Orange has been through experimental deployments of a cloud native core in its Pikeo network, using cloud-native core vendors such as Casa and HPE.

“Pikeo was easier because we had a more cloud native core. We didn’t have Nokia and Ericsson. Now we have real life and it’s more complicated. We know that if we begin deploying things as before then the journey to automation will be more complicated. The debate we have is to say, ‘How can we help you, Nokia and Ericsson, to define steps to move towards the cloud automation we have’. For them Nephio is a step towards Cloud Native deployment, but it is not reality today so we are working with them to define some stuff.”

So, if Pikeo was easier with more cloud native vendors, why move to Ericsson and Nokia for a commercial deployment?

“That’s a very good question. You know, we conducted the sourcing RFP on the core, and we took into account many parameters. It wasn’t only the cloud native readiness. It was the legacy systems, the compatibility – because we didn’t want for the five countries in Europe to make a huge change between what exists now and what exists in the future.

“We should not put all the blame on the partners because we have some strong commitments on Nephio and we have many R&D teams from Nokia working with us. For Sylva (Footnote 2) too we are working with partners, but we are aware it takes time. And we know we are also not ready for this shift of mindset because once we deploy it in the Kubernetes way even the roles and responsibilities between operations and the vendors is evolving and we’re working on that.

“I’m not sure either us or Nokia – I’m talking about Nokia because we can concentrate on Orange France but we have Ericsson in Belgium for example – are ready to set the roles and responsibilities between the partner, the operations of the country and us as Group for all the automation pipelines my teams are designing to deploy and configure and master the lifecycle management on all the CNFs.

“So it will take some time. We need to define the steps. We believe in Nephio – we think we can define those next steps thanks to Nephio – and we’ll see, we’re making some bets.”

Maadi says that Orange does want to move all its partners together to a Kubernetes based approach, with cloud native functions directly integrated into Silva.

“We have an ongoing RFP in MEA for 5G core network functions and for this we are very strict on the integration on Sylva, which is mandatory. And on some guidelines and topics regarding CI/CD we are very insistent because we know it will make our life miserable afterwards if we are not. So the idea is really to move all the ecosystem towards these cloud native functions even if it will take time.”

Gen AI and the Cloud

In the remainder of the interview, Roussia talks about Orange’s goal to do away with its front office in the NOC, the use of AI to enhance automation, its work to move data managent to Google Cloud and other public cloud platforms, and her view on Gen AI.

You can sign up to receive your copy of the magazine here.

 

  1. Nephio – A Linux Foundation progamme to deliver carrier-grade, simple, open, Kubernetes-based cloud native intent automation
    2. Sylva – Project Sylva is a telco-led programme in Linux Foundation Europe to create a new, open source production-grade Telco Cloud Stack