Orange's CTO and Senior Vice President of Orange Innovation Networks, Laurent Leboucher, spells out the benefits of achieving level 4 network autonomy by 2025.
Orange pursues service agility with Level 4 network autonomy
Like most communications service providers (CSPs), Orange is on a journey to transform much of its network to be cloud native and software defined. Laurent Leboucher, Group CTO and Senior Vice President of Orange Innovation Networks, sees automation as a key goal of the transformation, which will happen in phases.
First, Orange is applying practices like continuous integration and delivery (CI/CD) to automation of the network function lifecycle. The company has created “federated network integration factories” (within Orange affiliates plus a shared global service to pre-integrate common solutions) to implement a CI/CD pipeline with vendor partners. In parallel, for what Leboucher refers to as “in-life” network automation, Orange is leveraging data, AI and machine learning to find and resolve problems without human intervention in the live network. The final step is to have intent-based automation everywhere, which Leboucher says is “absolutely essential”. An early use case for it is service orchestration for 5G standalone slices, starting with static slices at first and then moving on to dynamic slices.
I spoke with Leboucher at length for our upcoming Digital Transformation Tracker report. In this first of a three-part series of articles drawn from our discussion, he explores Orange’s goal to reach Level 4 Autonomous Network by 2025. Part two will look at why intent-based management is so important. In part three, we’ll discuss the potential for generative AI and the need for cooperation among standards bodies and open-source groups. Our conversation has been edited for clarity.
Dawn Bushaus (DB) : Getting to a Level 4 Autonomous Network by 2025 is quite ambitious.
Laurent Leboucher (LL): Yes, it is, but we don't want to implement [autonomy] for everything. We want to focus on two processes: one is deployment, and the other is fault monitoring and maintenance. For these two processes, we want to reach Level 4, especially for the telco workloads, which will need to be significantly more agile than today.
I want to emphasize agility because very often we speak about efficiency of operations, but it’s agility and efficiency. Agility is extremely important to be able to deliver specific virtual networks – for instance, 5G slicing on 5G Standalone, and also global connectivity at scale for enterprise markets, leveraging SD-WAN, security, SASE, and so on. For these kinds of workloads, we need to be able to address customers in a different way and to be able to go much faster.
The other point to make is that we also need to become much more repeatable. If I have to deliver a mobile private network for a smart factory or a port, maybe the week after I will have to do the same for a railway company. We have started to do this in Spain – where we are now live [with 5G SA]. For the time being, this is something that we can still manage, but if each customer comes with their own SLA requirements, in terms of security, bandwidth, and latency … this is not easy and this is why we need to be extremely repeatable.
Security is also driving the need for change. For instance, if there is at some point a cyberattack that impacts the 5G core in a country, we need to patch the network almost immediately and be able to redeploy and restore it extremely fast if something goes wrong. We cannot take several weeks to do it. This can only be done if we have the full processes automated with the right images and artifacts in the proper repository.
So, change is key, both from a CI perspective and also from a deployment perspective, and then in-life, during the life of our operations. We need to be able to get all the correct information and be much more efficient, especially with anomaly detection. We can use AI and machine learning techniques to automate much more than we used to – and maybe in the future we’ll talk about generative AI for networks
So where are we? In general, we are starting with something which is close to Level 2, maybe in some cases a little bit less, especially with legacy systems either decommissioning them or making them fit to be automated for change or for monitoring. We are not yet at Level 3, and we will need to get there before going on to Level 4.
DB: What are the biggest challenges standing in the way of Level 4? I spoke with another TM Forum member recently who said he feels that standards, specifically a lack of collaboration among standards bodies and open-source groups, is perhaps the biggest challenge he is facing right now.
LL: You’re right, standards are a challenge. I would maybe not put standards first, however. The first one on my side would be culture: how to change the way people in network operations think. Because when you’ve worked using certain processes and practices for a long time in a given organization, it’s always difficult to consider moving to a different way of working. And this, by the way, is not only for us as a telco, but also for our ecosystem, especially the network vendors.
DB: What are some things that need to change?
LL: For instance, the way we manage risk in a telco environment. Traditionally, we are quite risk averse. And this is normal because we need to manage very critical infrastructure. At the same time, if we want to learn new practices fast, we need to be able to take some controlled risks.
Also linked with culture are skills and the lack of common ground, which are absolutely necessary if we want to succeed. For instance, if I ask, ‘What do we mean by intent-driven?’ and I have in the room a set of people coming from different operations teams, R&D teams, and maybe our partners … I will have different answers. There is a need to create the common ground and to define intent properly, but also to learn it.
Learning is also not only theoretical; you need to practice. If you are a doer, it’s much more powerful than if you are just learning in theory. So, I think we collectively need to take some risks and learn the hard way by putting it into practice.