Ray Le Maistre, TelecomTV (00:05):
So we're at the Linux Foundation's Open Networking and Edge Summit in London. I'm here with Arpit Joshipura. He is the senior VP and general manager for Networking and Edge at the Linux Foundation. Arpit, great to see you. Thanks very much for joining us.

Arpit Joshipura, The Linux Foundation (00:19):
Thank you for having me.

Ray Le Maistre, TelecomTV (00:21):
So in 2025 Arpit, what do telcos want from open source? I mean, what has changed say in the past five years?

Arpit Joshipura, The Linux Foundation (00:31):
Let me give you three things that have changed. The first one is open source networking has gone mainstream. And what I mean by that is from a decade ago since we launched projects like OpenDaylight, OPNFV, ONAP, Nephio, all of those projects, they have now moved to a deployment stage. So that's the first thing that has gone in, which means telcos have embraced open source and are using it in production network. So that's the first thing that has happened. The second thing is LF networking has emerged as the largest collaboration hub globally. And with the consolidation of other foundations like ONF, which now is part of Linux Foundation, OIF, Open Infra, which is now part of Linux Foundation, we now have an ability to provide more of a cohesive story to our end customers, like the telecom operators and make them more into an end-to-end solution.

(01:36):
So that's kind of the second big thing that has happened. And I think the third big thing from a trends perspective that we have seen is the lines between telecom networking, cloud networking, and enterprise networking have blurred because the common frameworks are Kubernetes the common framework. There are API layers coming on top that the enterprises can use, cloud and telcos are kind of sharing business as partnerships. So the lines are blurred. So we look at it more from a networking solution perspective than a market solution perspective. So I think those are the big three in the last five years.

Ray Le Maistre, TelecomTV (02:17):
And what are telcos actually looking for from open source?

Arpit Joshipura, The Linux Foundation (02:21):
I think they need flexibility of code and security of code. That's kind of the single most important reason why they believe open source is important. That gives them the speed of innovation and deployment. Normally it would take three years to deploy in a traditional proprietary environment, these things can happen in six weeks. So the time to deploy, time to fix bugs, if vulnerabilities come in, there's a whole process behind it. So that's kind of the most important thing. The second is cost. And it's not so much as the CapEx, but it's really the operational cost because interoperability is a given because most vendors participate in open source. The interoperability is already taken care of by these open source projects. So that's kind of the second big thing. And the third thing is it becomes future proof. Vendor agnostic lock-in is prevented. That's kind of what they always look for and they're still continuing to look for.

Ray Le Maistre, TelecomTV (03:24):
So the telco sector has been talking about cloud native for at least 10 years. Why does it feel like this sector is still dragging its feet? And what isn't critical evolution?

Arpit Joshipura, The Linux Foundation (03:38):
I would say it is not dragging. So I'll be just the first one to point out that it is not, and here's why I say that. In 2019, which is about six years ago, not ten six, I launched the term with my colleague at Mobile World Congress called CNFs Cloud Native Network Function. So Cloud native has two parts to the puzzle, the application part and the infrastructure part. The infrastructure part has come a long way. In fact, Linux Foundation Research just did a survey on the global community, and we found that over 73% of the global users and deployments are now cloud native based on the infrastructure side. So that is tremendous progress in just the last five years. As far as applications are concerned, that is dragging a little bit. So I agree with you. I just wanted to separate the two. The primary reason why it's dragging is there are physical network functions that are legacy.

(04:46):
There are virtual network functions that still have a reliability and a security requirement on it. And the newer functions, the CNFs are coming up and there's about, I would say they're at about 30% adoption right now. So there's a transition going on in the market where as CNFs come in, the VNFS are then terminated and moved over to CNFs. And so the call to action is for the application vendors and not the telecom service providers, but the vendors who support telecom applications to start and go with cloud native functions first versus just package a VNF in a container and call it CNF.

Ray Le Maistre, TelecomTV (05:30):
And do you think the telco sector from a cultural and operational and skillset perspective is ready for this adoption?

Arpit Joshipura, The Linux Foundation (05:39):
Yes, they are actually ready. And the reason they're ready is we have seen a lot of organizational changes within the telco where the three groups, the r and d groups, which is typically the CTO groups, the network operations group, and the IT group, they're now combining forces horizontally. So the best practices of CI/CD, GI Ops, DevOps, all of that is being implemented in the NetOps and in the CTO groups who participate in the open source. So we're seeing a lot of momentum in at least the top 30 carriers that are participating here.

Ray Le Maistre, TelecomTV (06:17):
So in terms of LF networking, what's the main focus at the moment? I mean, is it AI like the rest of the industry?

Arpit Joshipura, The Linux Foundation (06:25):
If I say no, then you will cut this interview short, right? But the answer is yes. And we at the Open networking and Edge Summit launched along with our platinum member Infosys, we launched two new AI projects. One is called Essedum and the other one is called Salus. Let me give a very short explanation of that, why this is important today, it's all about domain specific AI. So we're not talking about generic LLMs or generic models like ChatGPT, et cetera. The most important thing for telcos is how do I take advantage of all this AI applications in my network? So what they have to do is for every new AI application, they have to connect it with the models. They have to connect it with the data, they have to connect it with the infrastructure, they have to drive the governance around it, all the ML ops, AIOps, CI/CD, right?

(07:24):
There's a whole bunch of work they have to do for every AI application. Well, guess what? We need a framework to solve that. Where you bring in data sets, you bring in your models, but you don't have to do this glue logic every time. And that's what Project Essedum is providing provides that framework for AI in the networking domain. So we've seen tremendous response to that. And thanks to our partner Infosys, it's now part of LF networking as a project. So that's one. The second thing that is a more broader challenge is what is called responsible AI. And so we are announcing a project under the Linux Foundation called Salus, which is a shield, and it's for responsible AI. Essentially what we need to do is the ethical, the biases, the chatbot, there are so many responsibility constraints around the use of AI, and we want to make sure that it is actually coded up in a software project with seed code that people can confidently say that the use of AI is done well. So in general, we are calling it domain specific AI, and we are using that terminology for the telecom network operators, and we're seeing tremendous results on a lot of use cases at the LFN level.

Ray Le Maistre, TelecomTV (08:52):
Beyond AI, what else is the main focus of development right now at LF Networking?

Arpit Joshipura, The Linux Foundation (09:02):
So the first is continuing to focus on what the end users and our vendors need. So project maturity and deployment all the way to production, whether you call it virtual telecom, Orange, Verizon, China Mobile, AT&T, China Telecom, all of the above. So we are taking that into production, and what that means is integrating as many open source software pieces, Kubernetes, all the way up into what we are calling 5G Super Blueprint. So a blueprint, a super blueprint brings projects, hardware and software, including proprietary into a use case solution, which makes it very simple. In fact, in our survey we found 82% of the users want integration and interoperability to be done in an open environment, open collaboration environment. So we are in line with that. So that's bread and butter business as usual. The governing board pushes that number one. Number two is domain specific AI, which we talked about continuing to develop use cases for the networking, whether it's in the RAN or the edge or the core domain.

(10:25):
And then the third focus from a technology perspective is the last remaining puzzles for open source is the RAN. So out of the seven components of Open RAN, only four are open source. So the SMO, Xapps, RIC, O-Cloud, those are open source, CU, DU, RU is not. So with Linux Foundation, we are working with our partners to work a solution that can give true open source RAN components. So that's one focus. We're looking at the Edge cloud continuum as projects like all the LFN projects, silver projects, things like that allow for that level of deployment. And then the third focus is really on monetization. So projects like Camara that provide APIs to the network and allow enterprises to give value in terms of fraud protection or security detection or things like that. That's a big area of our focus from a Linux Foundation networking perspective.

Ray Le Maistre, TelecomTV (11:31):
Okay, so that's a fantastic set of developments ongoing there Arpit. That's going to keep you and all the members busy, I think for quite some time. Look forward to catching up again with you soon.

Arpit Joshipura, The Linux Foundation (11:42):
Thank you very much.