Red Hat aims to make the cloud infrastructure as easy to consume as possible. Azhar Sayeed, Senior Director, Global Telco Technical Development at Red Hat spoke with Telecom Review during the 5G MENA event about the key to building a multi-cloud strategy for telcos and how CSPs can maximize the cloud sustainably, among other insights.
Can you give the highlight of your keynote on ‘multi-cloud and the role of the hyperscalers’?
As part of the keynote, what I was really talking about is that hyperscalers have an important role to play with communication service providers, in terms of how they can augment the CSP portfolio with their capabilities. However, it's important to walk into that strategy consciously, rather than simply saying, “I'm going to partner with a hyperscaler”.
It's about creating your own strategy and leveraging the hyperscaler ecosystem to build what you want. With that end goal in mind, it allows you to think not about just one but multiple hyperscalers.
A hybrid cloud is important, but more importantly, multi-cloud is where you want to be. In that multi-cloud environment, you need to control your own destiny and build your strategy in providing services; not the other way around. You don't want them to control what you want to do.
In today’s cloud-centric era, which deployment models and strategies must CSPs focus on?
Incremental strategies have become really important. By building it incrementally, you start small, pick one use case, look at how to deliver that particular use case based on the hybrid cloud model, implement that, learn from it, and build the next one.
The edge is actually a good place to start because you may be able to push the public cloud provider into building that edge out with you. This will lead to lesser investments, in terms of CAPEX. When you start working with them with the same interest, you can potentially start there.
Co-creation models are also helpful and a win-win for both sides. There are two important points that you need to weigh: how quickly can you get the value out of what you're building and what's the opportunity cost of doing it.
Those are some strategies you can adopt if you are really keen on partnering with hyperscalers and leveraging their ecosystem in building services. However, you have to make sure that you are concerned about the pitfalls of doing this.
One of which is locking yourself into one ecosystem. The security posture and data sovereignty rules in terms of data storage are also the things that you need to control. You need to make sure that you have the right answers. And you're working with that cloud provider to get those answers.
If you're just using the public cloud as a consumer, it's completely different. If you really want to define your own strategy, don't let the cloud define you, and make sure that you deliver that capability.
Powered by 5G, how can telcos maximize the benefit of cloud sustainably?
With 5G right now, what people are more interested in is lighting up and building connectivity. Nobody's really utilizing the power of 5G except maybe a few providers around the globe. I think we still have some time to go in terms of the set of services that you can really see out of 5G.
At the moment, it's all about fixed wireless access (FWA) and high-speed broadband capability. Having said that, as you start to expand your footprint to provide high-speed services to consumers, you have to deploy more cells (small and micro) for in-building solutions. All of that consumes power. And as you start to go into massive MIMO antennas, we’re talking about 64x64 or 128x128 receivers. When you start to deploy that, you need an enormous amount of power in the cell site to make that happen.
Managing the power footprint also becomes incredibly important. That's where you can use the elastic benefits of the cloud. Keeping things static, where everything's laid out even when not in use, is not a successful strategy; it’s a waste of power and resources. In the cloud, you can create these services on demand and make them available. When there are no subscribers, you can start to scale back and shut them down.
Think about a stadium or a public place where you have a large number of users gathering for an event, and then after that, people disappear. You don't need the static capacity all the time. You can adjust based on the needs and the demand, optimizing not only capacity but power as well.
How do you see the telecom industry evolve further and how would Red Hat continue to fulfill its role in matching cloud services to global telcos?
Globally, we've seen telecom providers at various degrees in terms of cloud technology adoption. Some are just nascent and trying to just experiment their way around the cloud while others are far more advanced and very well-versed with what each cloud provider can give them and how to use it. The majority of that has got to do with the capabilities and the time needed to build upon them.
Red Hat’s goal is to make the cloud infrastructure, the software layer, as easy to consume as possible. Whether you're deploying that for RAN, edge, IoT, 5G core, or enterprise services, it needs to be consumable in a very easy way and provide all the capabilities as part of that infrastructure.
By using the Red Hat software as a particular infrastructure layer, you can also use it as a shim layer in a public cloud. If you develop and/or deploy an application under a private environment, by using the same Red Hat software layer in the public cloud, you will have the same exact interface, experience, and management capabilities.
We are enhancing the platform because, to illustrate, RAN requirements are very different than running a web application on the cloud. For RAN, you need to worry about timing synchronization, hardware acceleration, performance, and power while in a web application, it's a simple HTTP.
Thus, we have an all-inclusive cloud abstraction layer that caters to different requirements by being automated, highly available, and consistent.