Meet the speakers

Andrea Bozzoni (00:01):
Welcome to the Microsoft Azure Red Hat® OpenShift® video series. I'm Andrea Bozzoni, from the Red Hat Managed OpenShift Black Belt Team. And here with me is Roy.

Roy de Milde (00:11):
So I'm Roy. I work for Microsoft in the Global Black Belt Team based out of the Netherlands. And today we're going to talk about use cases, innovative use cases. When we start to think about what Azure brings to the table, what OpenShift brings to the table, and all the native integration that we see, there are actually some really cool use cases that we can think about. So what I always like to do is dream big, and then even a little bit bigger. So if I can ask you, Andrea, if you dream big, what do you think about and what do you see as innovative use cases?

Innovative use cases in Azure Red Hat OpenShift

Andrea Bozzoni (00:43):
Yes. When I dream big, I dream for the customer. And if I think for the customer, what the customer really needs today, they need power, because they want to deliver solutions like artificial intelligence, deep learning, machine learning, and scientific simulation. On this side, Azure gives us the availability of the GPUs, V-Powered virtual machines. And, on the other side, ARO has the capability to use this virtual machine in a simple way, as worker nodes in an ARO cluster. So this allows us to deliver this power when the customer needs it, in a flexible way, because you can set up in an ARO cluster, add this virtual machine as a machine set, and use it in a flexible way. In this way, all these things are convenient together and give the flexibility the customers are searching for today, to deliver this powerful solution. Shall I go to the Lightboard and quickly explain to you how to create an ARO cluster with a powered GPU virtual machine? 

How to create an Azure Red Hat OpenShift cluster

So let's see how to set up an ARO cluster with a GPU's machine set. So first of all, imagine that we have an ARO cluster with three master nodes. And the cluster has three worker nodes set up.

(02:08):
We create different GPU vector machines, like for example, Tesla V100 or P100, but there are many more available. And then with three steps, we are able to deliver this machine set. To add this set to our cluster, first of all, we need to add the NVIDIA GPUs, operator, that takes care of all the complex stuff to install drivers, Kubernetes plugin on all these machines. The second step is to install the node feature discovery operator. That allows ARO to discover all the GPUs on this virtual machine. And in this way, every workload can be targeted on this. And the last step is to deliver the workloads on this node. And that's it.

Roy de Milde (04:01):
Thank you. That was amazing. Thanks. It's really cool to see.

Andrea Bozzoni (04:03):
Yeah, it's really amazing. And if you would like to dream big, what do you think about?

How to build worldwide clusters with Azure Red Hat OpenShift

Roy de Milde (04:08):
Sure, sure. So the way I think about it is, we have a worldwide presence with Azure. Azure comes with more than 60-plus regions all across the world. So if you combine that worldwide reach in combination with Azure Red Hat OpenShift, and then capabilities on top of that, you can build all kinds of worldwide clusters. So let's for example, take Red Hat AMQ, the solution by Red Hat, a messaging broker solution. So what we can do over here, if we combine all the three flavors, so Azure Red Hat OpenShift in combination with AMQ, we can actually build out a full-fledged global messaging platform. But shall I quickly draw that for you and tell you how that potentially would work?

Andrea Bozzoni (04:53):
Yes, could be really interesting.

Roy de Milde (04:54):
Awesome.

(04:58)
So I was talking about a global messaging platform. So the first thing that I'm going to draw over here is the Earth, right? We have 60-plus Azure regions. So if you look at one of those regions, we have an Azure environment. So I draw one over here, but I want a second one and a third one. So I draw one over here and over here. And the cool thing is that with these Azure environments, I can choose the region where it needs to be deployed. And we have 60-plus, right? So this one is going to go over here, this one on the south, this one over here.

(05:46)
Just as an example. Now within those Azure regions, we're actually going to spin up an ARO environment. Easily done. You can use whatever tool you want to use and spin it up really fast. Really quick, I'm just going to name one “ARO”. Cool. But on top of that “ARO”, we can now also easily deploy, for example, AMQ. So we're going to deploy a custom workload. We call that “AMQ”, do the same over here obviously, but just pointing it to a different region, and over here as well. Now, super cool. I have three clusters, three AMQ instances spread across the world. So we need to make sure that these clusters can interact, and we can do that over the Azure backbone. But then we can actually also do some cool things over here and point this all to a traffic manager. And there you have it, a full-fledged integrated global messaging platform powered by Azure, Azure Red Hat OpenShift, and in this scenario, AMQ as well. I I hope you enjoyed that. 

Andrea Bozzoni (07:12):
Yes. A lot.

Roy de Milde (07:13):
Cool. Any, any other use case that is on top of mind for you?

How to use Azure Red Hat OpenShift to deliver solutions quickly

Andrea Bozzoni (07:16):
Yes. I think that the customer now also needs some tool services and application frameworks to deliver solutions fast, and to answer requests that businesses, every day, are asking them. They are putting them under pressure to ask for new functionalities on their application. So there’s a need for this IBM developer… software packages and tools, like all of the IBM Cloud Paks, that allow customers to use this kind, that they deliver to the customer. Services, application framework, and tools that allow them to be faster in the delivery solution. And this kind of IBM Cloud Paks can be delivered on top of ARO. For example, if you need to deliver an artificial intelligence or analytic solution, use IBM Cloud Pak for data… gives you all the tools, the machine learning tools, the machine learning models, to allow you to quickly deliver this kind of solution on an ARO cluster. And in this way, the customer can concentrate on the business value and not on the technical things that are behind this. So it gives them the possibility to quickly deliver the solution that they need to release for the business.

Roy de Milde (08:26):
Super cool. So what I'm hearing, ease of Azure obviously. The ease of deploying Azure Red Hat OpenShift, and now with the IBM Cloud Paks, we have full-fledged integrated solutions that we can just deploy on top of that, so that the customer can focus on delivering business value. Yeah, that's amazing. I would love to get my hands on that and play around with it. Yeah. Cool. 

Low-latency e-commerce and Azure Red Hat OpenShift

There is one other thing that I was thinking about, and that is about a low-latency e-commerce platform. So what we know in the retail business and with an e-commerce platform is that you wanna make sure that your customers have the shortest way to your platform when they're going to buy a product. For example, because we also know, based on research, that if it takes longer than three, four seconds, they're probably gonna go away…

(09:12)
…and go to maybe the competition to buy that product, right? So what we are actually seeing with Azure Red Hat OpenShift, and an e-commerce platform, on top of that, and with all the advanced capabilities around networking, we can build those low-latency e-commerce platforms. So think about Azure Front Door with the content delivery network, but also application acceleration, advanced network acceleration. And if you combine all those different things together, now you get a global scale, low-latency e-commerce platform guaranteeing you that wherever your customer is around the world, they will get the shortest path to your platform when buying your services or product. Right? So I think Andrea and me, if we sit down for another night and we're going to think about it, we can think about many more, right?

Andrea Bozzoni (10:03):
Yes, there is a huge variety of workloads that can deliver on our cluster.

Roy de Milde (10:08):
For sure. I fully agree. That was it for this video. I hope you liked it. We would like to challenge you to dream big, as well. And if you have any questions,and you want to dream big together with Red Hat and Microsoft, feel free to reach out to your Microsoft or Red Hat counterpart. We are here to support you and help you, and we hope that you enjoy the other videos in this series as well. Thank you.

Andrea Bozzoni (10:32):
Thank you. Bye.