Q&A: Telcos love open source, but still dither about ROI

open source telco
Peter Man, GM at Red Hat Hong Kong, Taiwan and Macau

Telcos have a reputation for being resistant to open-source technologies – or at least being unable to fully understand how they can leverage it. The reality is a little more complicated. In this interview, Peter Man, GM at Red Hat Hong Kong, Taiwan and Macau, outlines to Disruptive.Asia editor-in-chief John C Tanner the state of play with open source in the telco world.

Disruptive.Asia: It seems that telcos are getting mixed messages about open-source – they like the idea of not being locked into a single vendor environment, but even the network equipment providers (NEPs) adopting open-source are still pushing single-vendor scenarios as an end-to-end play in terms of maximum efficiency, O&M, tech support etc. Is that confusing operator CXOs?

Peter Man: The traditional NEPs have been working with these telcos for 20 years, at least, so they know them very well, and if I were them, I would tell them the same story. But if you look at what Red Hat is doing, every one of those NEPs are our partners. And especially on the NFV/NFVi layers, they are all using open source as well.

Of course, the beauty of open source is you don’t get locked in by the vendor. The difference between open source and proprietary software is that with proprietary, you really cannot move away very easily. With open source, we only sell subscriptions, which is the support service. When you get the source code, you get everything else for free. So technically, you don’t need our support – you can download all our source code and compile the software yourself. The only difference is, we will not support you because you’re not paying us for support. If you are able to support the whole thing yourself, you can do it. I believe it is already happening in some of the mega-sized international telcos – in the long run, they will support some of the open source components themselves. It depends on the situation.

So NEP players still want to have control and have a have a big share of what’s ahead. But at the same time, customers are not stupid. They know they don’t want to get locked in for another 10 or 20 years. They know that open source unlocks the value of next-generation services to something that’s more than just moving data around. I always use the example of a bank – if the purpose of a bank is just moving money around, the margin is very slim, because all the other banks can do the same thing. So the banks are targeting services that they can provide as a basic function, and the premium services that people are willing to pay a fee for.

For telcos, it’s very similar. If you’re just moving data around, everyone can do that – it’s not rocket science. It’s the same with 5G – I just read that Apple will not put out a 5G iPhone until 2020. The reason is because for most apps and services, 4G is enough – even for streaming movies. Are you willing to pay a 30% to 50% premium for 5G just because it’s faster when your apps already work on 4G? It’s like your water tap in your house – who would pay the water company an extra 15% for a faster flow of water?

So the real point is how fast the service provider can roll out new services so people can enjoy the benefits of 5G. For that, you need to build a platform to provide premium services. That’s the part telco players are interested in.

And that’s the reason you need an open standard to decouple the software layer from the hardware. You cannot rely on one single vendor to provide you all those premium services because you need to follow their stack – if your partners are not on that stack and they cannot communicate with you, you will have a hard time bringing in other people’s services.

The other issue is time to market. Going back to the example of banks, they have a lot of competition today from the virtual banks, P2P lending services, who are getting their services out quickly. The same thing is happening with telcos.

Telcos get that part. The difficult part, which we are working on, is to make sure they can feel comfortable with this new way of engagement with different technology partners. That’s the hardest part – not just for telcos, but for other industries as well.

What exactly makes them uncomfortable?

They’re used to the setup that an NEP will come in and support everything, so they have one number to call, and one neck to choke. At the business level – the CFO or the CEO level –people are starting to get it. But it’s more like the culture, the mindset and all these layers of the business – it’s more on the working level or the mid-management level that they need to feel more comfortable with it.

Telcos understand the benefits of open source, but they’re struggling on how to do it. When they make an investment to go to an open platform, they need to justify the ROI. So it’s more like a wait-and-see game, especially for the players in Hong Kong, to see how things evolve, and which way they should go.

But they’ve been playing wait-and-see for several years now – how much longer can they wait?

After they do the investment, they need to see the ROI in the next three to five years. That’s the most difficult part because you don’t know what will pay off quickly – it may be a long journey. So even if they know what’s going on in the market, they want to make sure that they see the end of the tunnel.

What is Red Hat doing to help them along that journey?

Telcos in Hong Kong are very interested in putting next-generation services and applications in container technologies. They’re moving away from the traditional scale-up architectures, because container is a lot more flexible, especially for their kind of services, because most of the major players in Hong Kong have like half a billion subscribers already. So if you launch new services and demand grows very quickly, you don’t want a traditional infrastructure to power those applications. You want a flexible way that you can scale out on your own cloud or even on the cloud of third parties. For traditional services, telcos would engineer their networks to account for peak demand that was predictable, but nowadays services come and go, so traffic isn’t so predictable anymore. You could maybe lock in customers to subscribe to your service for a long period time, but a lot of customers are spoiled by AWS, where you can subscribe to the service today and cancel it by tomorrow.

So you need to have an infrastructure that is designed to handle this kind of come-and-go, so you can scale up and scale back when you need to. That’s how public cloud works, that’s how container works, how microservices work. All of these are open source technologies – Kubernetes and Linux. So I think all these moving pieces come together and make Red Hat a very obvious component in the next-generation architecture.

Now that Red Hat is being acquired by IBM, what will they bring to the table in that regard?

One thing I can tell you is that combining with IBM, we definitely see that we can go to market much quicker compared to doing it all by ourselves. IBM still has a lot of footprints in many large enterprises – that part will definitely help us. IBM will keep Red Hat running independently after the acquisition so we can continue to be the Switzerland for everything. So we will continue to pursue our white space customers, and we will help the IBM cloud story become more solid. And as our CEOs have said, we will be the biggest hybrid cloud player in the world. That’s all I can answer on that one.

Be the first to comment

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.