Orange on cloud native—‘It’s about agility, scalability, resource efficiency, cost optimization’ 

New paradigms around infrastructure automation and software development are guiding the cloud-native reinvention of Orange

Change is afoot in the telecom industry as the move to cloud-native operating and technological principles push and pull operators into new paradigms. With a renewed sense of conviction about the network as a platform for innovation—not just a utility-like network of dumb pipes—multi-national communications service provider Orange is undertaking a massive reinvention, from telco to techco if you like, that’s all about “agility, scalability, resource efficiency, [and] cost optimization,” as Philippe Ensarguet, vice president of software engineering, explained during the recent Telco Cloud and Edge Forum, available on demand here

Citing a renewed interest in leveraging network APIs via consortia like Linux Foundation’s CAMARA and GSMA’s Open Gateway Initiative, Ensarguet said, “We are at a moment where telecom operators basically are more and more convinced about the value of their network…The pjatformization model is something that is super, super important.” And in this shift, he said there are lots of lessons to be learned from hyperscalers and enterprise IT, sectors that he reckoned are at least five years ahead of telecoms. 

“It’s happening,” though, Ensarguet said. “The software transformation, the cloud transformation, the data transformation…it is happening right now. It means we are entering into new paradigms to implement our infrastructure. And I want to talk about the move from the very closed and proprietary, I would say, vertical model toward the horizontal model that is heavily…cloud-based. And driving this shift definitively for us opens new ways to operate our services and to be more efficient.” 

The big picture here, in terms of guiding principles in the shift to cloud-native telecom networking, was laid out in a manifesto published last year by the Next Generation Mobile Networks (NGMN) Alliance with contributions from Bell Canada, BT, Chunghwa Telecom, Deutsche Telekom, Orange, Telia, Telus, Turkcell and Vodafone. 

The cloud-native guiding principles as articulated in that paper are as follows: 

  • Decoupled infrastructure and application lifecycles over vertical monoliths  
  • API first over manual provisioning of network resources  
  • Declarative and intent-based automation over imperative workflows  
  • GitOps principles over traditional network operations practices 
  • Unified Kubernetes (or the like) resource consumption patterns over domain-specific resource controllers
  •  Unified Kubernetes (or the like) closed-loop reconciliation patterns over vendor-specific element management practices  
  • Interoperability by well-defined certification processes over vendor-specific optimization. 

Borrowing cloud-native best practices from hyperscalers and IT

While this is something of a sea change for many operators, Ensarguet pointed out that it isn’t like the goal is to reinvent the wheel. “We all know that basically the operating model for cloud-native is GitOps. So when you are using GitOps, basically you are using technology and components and projects that are coming from those [hyperscaler/IT] ecosystems.” The big changes, he said, are around intent-based, automated operations of a disaggregated network, and the move to a distributed, service-oriented architecture.

Beyond this technological overhaul, Ensarguet stressed the importance of workforce and organization mindset shifts; essentially a departure from business as usual. For instance, he said, the idea of service reliability engineering, borrowed from Google, has been adapted for telecoms to network reliability engineering. And, of course, the rise of artificial intelligence, both classical and generative, sets up further necessary skill adjustments. 

“When you’re learning, you are learning about the tech, you are learning about the methodology, but you are also learning about also how the skills need to be transformed,” he said. We are not in front of a simple evolution…We are in front of what I’m calling a revolution because it’s about, I would say, the culture…It’s about bringing a way to measure where the products are, where the services are. So it’s about assessment, gap analysis, and then it’s about educating, bringing awareness. We have an intense program of upskilling and reskilling.” 

In an AI world, experts are hard to find, expensive and very much in demand. This reinforces the need to upskill/reskill; when the people you need are hard to get, you have to count on the people you already have, he said. 

Comments are closed.