When it comes to research and innovation in the network space, there are a lot more bigger and better funded sources than the internal R&D departments of network operators.
Vendors themselves, of course, eventually rise or fall on their ability to innovate and efficiently integrate new technology. Then there’s academia and all the work going on in university labs over the world, often industry-supported. In a more commercial vein, there are dedicated R&D houses carrying out applied research and exploration. Then there’s VC capital, with its funding of start-up innovation and development. Increasingly, in some jurisdictions, government money is being funnelled to research projects in strategic areas of national interest, such as security, public safety and sovereign AI. All of this has been placed in context in recent years by the big tech companies that have pushed out cloud-native and now AI ops paradigms, and the billions going into new AI models and hardware platforms.
Within this landscape most major operators, of course, carry their own long-standing R&D funcitons, tasked with understanding upcoming technologies, contributing to standards development and securing intellectual property for their businesses, as well as fostering and enabling innovation in a wider sense within the business. That’s a wide area of operation, so how does a network operator approach research and innovation in the network space, within the evolving network paradigm?
Éric Hardouin, pictured at Orange’s 2021 Research Exhibition.
To find out more about one operator’s approach, TMN spoke to Eric Hardouin who is VP, Networks and Infrastructures Research, Orange. His division has about the equivalent of 100 full staff working in the network and infrastructure domain, which covers Orange’s fixed and mobile networks from access, transmission and core, as well as what it calls home LAN technologies. In scale, that’s way fewer people than would work in the R&D function of just one domain within the large vendors that supply Orange.
But it doesn’t mean the department is irrelevant in shaping the commercialisation of technology within the business. Far from it, Hardouin would argue.
In the future telecom landscape – based on software, cloud and AI – having this force of research to identify and design new, disruptive solutions is clearly an asset for us.
Overall, Orange Research is structured into five domains; AI for network operations, AI for customer relations, Cyber security, Sustainability and Hardoiun’s own Networks & Infrastructures domain.
Hardouin breaks down the department’s contribution into four main areas. First of all is to establish a vision of how a technology is going to evolve, and how Orange would like it to evolve to best suit its business, and to promote that into the ecosystem. A good recent example is Orange’s taking a lead on the sustainability aspects of 6G.
“I wouldn’t claim we were at the origin of the strong sustainability push but we played a role in it. At the first wireless 6G summit in 2019 we were one of only a handful of companies to talk really strongly about sustainability, and now this has become one of the key pillars that everyone expects,” Hardouin says.
The second function is to solve hard problems for the business functions. An example here is the development and deployment of tools to help technicians deploy and operate access networks.
Another major role is to create skills that may be needed at some point by operational colleagues. Hardouin points to the engineering of the Orange Quantum Defender network, which provides quantum safe services in the paris area using QKD and post-quantum crytography. Orange started to work on PQC in 2017 and QKD in 2019, and it was the research team which brought forward the engineering which is the backbone of the network. This ability to generate technology is also reflected by orange’s patent portfolio: Hardouin points out that thanks to its research, Orange holds 600 patents within the portfolio of a typical 5G smartphone.
A fourth aspect is to use the research department’s know how to evaluate technology proposed to the operator by suppliers, so that research experts are involved in evaluating vendor solutions.
Becoming more business-focussed
One key switch up in the last two years has been to bring the research units closed to the business units within the operator.
“Until two years ago we were kind of autonomous,” Hardouin says. “We were deciding where to invest resources based on what we were seeing in research conferences, and our own ideas, and making decisions based on research.”
Now the department tries to engage more tightly with business stakeholders, getting feedback on what the research teams are working on. This aligns the value chain from research to business, with a shared vision and focussed resources.
Hardouin: “It’s part of the strong move from Orange Innovation, within which Research sits, to prioritise topics of value to customers and to the operator.
“The goal is to make sure we have a shared vision about the challenges, and the roadmap we should deliver. So we make sure we streamline all the innovation, from research to exploration, and do that with strong links to the business units, so we are all aligned on what is important and where we should invest our resource.”
One result has been the integration of research with exploration, which is the downstream part of research. For example, all the activities in research and exploration of Non Terrestrial Networks (NTN) are under Hardouin’s responsibility.
Hardouin lists another good example – the development of a network deployment tool for the network operations team. This started life as a complex research project being worked on by a PhD student. When research introduced the topic to the head of the relevant business area, the feedback was that it was far too complicated and to come back with a simple MVP that could be put into trials. That happened, and now things have developed again so that a more complex v2.0 solution is also in field trials.
Hardouin says there are other similar examples that he is unable to disclose.
Taking a business-focussed approach will also stop research on science projects that look unlikely to deliver a business benefit. For example, Orange stopped exploring THz research as a candidate technology for 6G back in 2020, when it judged that it just would not be viable commercially, requiring a deployment grid beyond the financial capabilities of most operators, with little likely payback.
As Hardouin assesses it: “THz will cost so much money that it is not compatible for macro. Or it could be for indoor, but we have not even deployed mmWave there [yet], so why should we care about it?”
We don’t have the means to deliver solutions for everything, and we may not be the most relevant for doing that, so we try to work where we think there is a lack of investment in the ecosystem compared to what we think is important,
Existing in the innovation ecosystem
The 6G THz issue raises one interesting research angle, which is how Orange operates within the wider innovation environment within network R&D.
Orange does have bilateral and collaborative programmes with vendors, which it can use to deliver its priorities to the vendors. And it can also use its skills to understand the details of what vendors are proposing – for example understanding the implications of an algorithm in terms of energy or potential hardware cost.
Aside from that, Orange can use its research resources to prioritise areas it thinks are undervalued by industry.
“We don’t have the means to deliver solutions for everything, and we may not be the most relevant for doing that, so we try to work where we think there is a lack of investment in the ecosystem compared to what we think is important,” Hardouin says.
And Hardouin points out that Orange has been able to partner with some of these major companies, for instance with OpenAI and Mistral, with whom it has a bilateral research partnership. This has been helped by the forward looking mindset of Nicolas Demassiuex, who before his retirement as head of Orange Labs Research saw that there was a need for Orange to invest in AI research.
This research aligns with the development of Orange’s networks to become increasingly cloud-native, and operated along cloud ops principles. That means Orange will have more opportunity to inject its own algorithms into network operations, led by its research and partnerships in AI.
“In the end,” Hardouin says, “in the future telecom landscape – based on software, cloud and AI – having this force of research to identify and design new, disruptive solutions is clearly an asset for us.”
