The Multi-Cluster OpenShift Mindset with Rodolfo Casás

Pete Wright:
Hello everybody, and welcome to Trilio Insights on TruStory FM. I’m Pete Wright. Today we’re diving into the complex world of multi cluster OpenShift, a linchpin in the realm of modern application deployments. As businesses scale and the demand for resilient, cost-effective, and efficient operations crescendos, managing multiple Kubernetes clusters isn’t just a challenge, it’s an art.
Rodolfo Casás is our senior solutions architect here at Trilio, and joins me this week to unravel the intricacies of multi cluster OpenShift, explore the tools that conduct these distributed environments, and share tales of transformation from the organizations hitting all the right notes. Rodolfo, welcome.

Rodolfo Casás:
Hi, Pete. How are you doing?

Pete Wright:
I’m doing very well. I’m very excited for this Rodolfo. What listeners don’t know is that you sent me an outline a while ago, and then just before we pressed the red button said, “Okay, Pete, I have things to talk about. Get ready.” That’s effectively what I heard you say. So I’m hoping we could start with a kickoff on the concept. We need to talk about multi cluster OpenShift and why it’s important. And I know as an experienced trainer, you’re going to be able to break this down in such a way that I and our listeners will be able to understand and adapt. Where would you like to begin?

Rodolfo Casás:
Yes. So yeah, OpenShift, many people has come in to this world of OpenShift and they don’t know the whole story. That’s why I wanted to more or less talk a little story. The world of IT is changing. So internet came, and then small companies, small online retail companies, online stores, streaming services, Netflix, the Spotifys, they came to this world. And then suddenly a small company could become really a big company with millions of users in days or weeks. So that brings issues into the world of IT. So some of the challenges of these companies is modernization of systems and adapt to new industry trends and new legislation, a global scale. Some of these, the Ubers and the Starbucks, they work at a global scale. So how do you handle that with legacy tools, and legacy environments, legacy infrastructures? It’s really difficult.
So I would like to start with some new methodologies and what they call digital transformation, and with the help of cloud computing, containers, automation. As a [inaudible 00:02:52] instructor, I’ve seen sometimes companies before, they had a lot of people managing a small number of servers, but now you can find maybe three guys managing an incredible infrastructure, and it’s all thanks to these tools like OpenStack, Kubernetes, automation, containerization, programmability, software defined networks. It’s all very complex, but at the end of the day, it allows you to achieve these goals.

Pete Wright:
Okay. So I know you have broken down these concepts into three main, I’ll call them buckets, for our conversation. What is the first requirement to be able to transform from this nascent organization to global scale?

Rodolfo Casás:
Yes, so I think it was AWS, started with this concept of public cloud computing where they thought, okay, we have a lot of computing resources available for Black Friday and Christmas, and what do we do the rest of this year with it? We have those servers smoking cigars as I say. So they decided to open up AWS services, and you could hire temporarily an EC2 instance or just use some storage from them in the cloud. At the time when I was looking at that, it was like, “Wow, this is magic. You can have a server really for a short period of time, depending on the size, you paid more money or less money. And you could put some data over there. And it was really cheap and very resilient with a lot of nines of resilience.”
That’s the public cloud, and that has evolved from just a couple of services in AWS to, I think there’re now more than 150. And now it’s not just AWS, we have Azure, we have Google Cloud, IBM Cloud, Oracle Cloud infrastructure. There are many public clouds.
Now the thing is not all companies and workloads are suitable for migration to a public cloud. The public cloud gives us many advantages, but there are, as I said, not all companies can do that. For example, some critical workloads running in the data centers, they will never will go into the public cloud.

Pete Wright:
Like what, can you give me an example? What don’t you want to send? What services don’t you want to send to the public cloud?

Rodolfo Casás:
Well, if you talk to a public cloud representative or sales guy, they will tell you anything can be run on the public cloud. But then some of the people will say, “No, my data has to be in my data center,” for example, of maybe for some regulatory or compliance reasons, some regulatory requirements from some countries. They, “No, no, this, you cannot run personal customer health data in the public cloud,” for example. Or maybe other customers cannot run because they have to rearchitect their applications to run in the public cloud and they’re not ready yet. Or maybe it’s too expensive, it is cheaper to run them in the private cloud. Actually, I’ve seen enterprises going to the public cloud and then they get surprised because, “Oh, this is more expensive than what I thought. Can we try to go back, at least some of our service to the private cloud?”. I’ve seen that.
And then there’s third model. There’s a third model. Why don’t we run some workloads in the private cloud, like OpenStack or Kubernetes, and then we can run some other workloads in the public cloud? Or some of the customers, what they’re doing, is when they need more computing resources, or more storage, or certain services from the public cloud, they run some workloads in the public cloud for that reason.
There are some concepts that define a cloud like cell service, for example, that’s really important, or scalability, so you can scale up and scale down your applications. Companies like new startups, they can start very small and then grow very easily because the hardware resources are already there. You have that in the public cloud.
Now, why is scalability useful for me in the private cloud? Let’s say I have a private cloud like OpenStack or Kubernetes in my data center, it is very useful for me if I have the possibility, which is already available in tools like Kubernetes and OpenStack, you can run auto scaling tools. So the applications, let’s say you have 50 applications, so the applications that need more resources, they will get more resources in real time, and if they need less resources, they will use less resources. So the cloud concept is very useful in any private, public, or hybrid cloud.

Pete Wright:
Okay. Let’s talk about containerization.

Rodolfo Casás:
Containerization. Whenever I explain containerization, I just have to bring up my history. I start, one of my first jobs was working in Fujitsu with big spark prime power servers, selling this prime power with hundreds of CPUs to big banks and big telco companies. So the goal was to get the most expensive hardware with a very good operating system so it was very reliable and it never fails. It’s completely reliable.
The whole concept of the cloud is not like that. It’s assumed everything is going to fail at some point. So what I need is some kind of redundancy in every component of my architecture, and that allows me to bring cheap hardware to all these concepts. So that’s it.
And then in that world, we could isolate workloads from another using, for example, system boards or CPUs. Then it came virtualization. So now I could run [inaudible 00:09:02] machines inside of a physical server. And then the next level of isolation is containers. So I can run my Apache web server, or my engine X web server, or mySQL just using a process. I don’t need a whole operating system to run one workload, so I can have a lot more of applications using the same amount of computing resources. That is how I explain containerization.
Now, on top of that, you get a lot of other benefits and features, but that for me, that’s the, let’s go. Let’s save money because money moves everything. You know what I mean?

Pete Wright:
Yeah, right. Well, and that leads to I think the next money saving operation, which is automation. That’s the third jewel in the crown.

Rodolfo Casás:
Exactly, exactly. How do we manage all this impressive amount of services, networks, the software networks, containers, applications, users, systems, clusters? There’s no way we can do that without automation. Automation helps you to, well automate stuff so you don’t get… Sorry. You try to avoid human error and to speed up processes. In the old times, you get, “Oh Rodolfo, could you create these five VMs with this operating system and this configuration?”. They’re all the same, right? But I tell you, at the end of the week, I get five servers, none of them are identical because they’re all built by hand.
Right now with automation, you can do that automatically very fast, and they’re all the same because they’re infrastructure’s code. So you just put it in the code, you test it, and then it’s going to be the same forever. And you can run that automation in any cloud against one host or a thousand hosts, against one Kubernetes cluster or a thousand. It’s just whenever I ask any of our prospects of customers, “Hey, do you use automation tools?”. Absolutely all of them are using it or in the way of implementing it.

Pete Wright:
Okay. So those are the big three, right? Our cloud, our containerization, our automation?

Rodolfo Casás:
Uh-huh.

Pete Wright:
And that leads us to what? Are we coming back around to your principle thesis on celebrating OpenShift together?

Rodolfo Casás:
Yeah, multi-class OpenShift, OpenShift anywhere. So let’s say I’m Red Hat. If my customer wants to run applications in AWS, Azure, Google, on-prem in the edge, they built a product, OpenShift, that can adapt to any of these environments. And you can run OpenShift anywhere. And you manage OpenShift the same way everywhere. I don’t know, I think the last time I checked there are like 60 conformant CMCF distributions, but they are all managed in a different way. It’s like distros. You have Ubuntu, you have Debian, you have Arch Linux, you have Red Hat. They’re all managed in a similar but a different way. And the same thing happens with Kubernetes.
Now you have the power, and you have the flexibility, but how do you handle, you reduce that complexity because it’s becoming more and more complex every day? Yesterday I got a message, an email from the Linux Foundation saying, “Hey, previously if you passed an exam, a certification exam, it was valid for three years. But now as the amount of changes happen more and more often, we’ll reduce that to two years. So if you certify with us, your certification will only be valid for two years, you’ll have to recertify. Yeah. So that is a proof, a testimony of how quick this is changing. It’s just an incredible spin.

Pete Wright:
Give me a sense of when organizations, let’s say I’m CTO, right? Let’s say I’m a chief engineer. How am I going to run headlong into the constraints that require me to consider how I’m managing across OpenShift? What is the practical consideration of constraints that I need to look into? I’m sure you’re going to come back to saving money.

Rodolfo Casás:
Yeah. What I’m trying to say is, well, OpenShift is another distribution among others. Now, given the capabilities of OpenShift, their multi-class architecture, the different software products that they have embedded inside OpenShift, we truly consider a market leader. It has a big market there. And it is just because of all these reasons, it has security integrated, it has policy management with ACM, it’s all integrated. It has automation without some [inaudible 00:14:09] automation platform.
So we at Trilio, we thought this is… We’ve been partners of Red Hat for many years and we have adapted our product to have a great integration with Trilio. Now, when one of the struggles that a CTO will have when deciding what distributions I run, where do I run my [inaudible 00:14:34] distribution is for example, management. And then it’s not the same if you have different distributions in different clouds and also on-prem, they will have a different way of management usually, and that adds complexity instead of reducing complexity.
With OpenShift, you can manage all the same everywhere. And they have a [inaudible 00:14:56] manager called Advanced Cluster Manager where you can base your configuration in policies. And that’s where we, for example, decided, Hey, we need to integrate Trilio with this Advanced Cluster Manager policies. For me personally, whenever I present, I introduce a product to anyone, I say, “Listen, you know what? If you’re using OpenShift, if you have Advanced Cluster Manager, this is the most simple way to manage your backup strategy because you just configure a couple of JAMA files, and whenever you have add a cluster to your ACM hub, it will install our Trilio operator, and then it will start taking backups automatically. It’s all like magic, okay? It’s just enforced.”
It’s just, how did we go from spark servers where you have to do all manually and be extremely careful to someplace where you just deploy a couple of text files and it takes care of everything? For me, this is just crazy, this is just unbelievable.
And that is all the thing I wanted to present.

Pete Wright:
I’m here for it. Can you just give me a state… This is the thing I really love thinking about, which is our place in the industry, because what you’ve described in your own words as magic, this feels like a thing that we should be seeing a lot of adoption around. Do you get a sense that the industry is looking kindly on this? Are we moving in this direction of maturity here?

Rodolfo Casás:
Yes. Definitely, yes. Because what happens, a customer starts having one cluster, two clusters, and then QA clusters. I run temporary workloads on a cluster, but then I cannot sat down that cluster because I want to do some testing afterwards. And then another department has some development clusters or telcos. They have hundreds of clusters distributed geographically. So we can go using Kubernetes or OpenShift, we can go from just two clusters, which there are a lot of companies just using two for redundancy, for disaster recovery. But there are many companies already having hundreds of cluster, like banks, financial institutions, health companies especially, especially telecommunications companies. Now all these 5G thing, and the open run, it’s bringing a lot of clusters to the table.
So how do we manage all that? So you need a manager, you need automation, you need tools that can, as I said, reduce complexity, reduce complexity, make it simple. I always say that, “We give you the power and we help you to reduce the complexity of your architecture.”

Pete Wright:
Outstanding. Well, I mean, if you haven’t given us a testament to reducing complexity while increasing scalability, I don’t know how else we can make that point. Thank you so much, Rodolfo.

Rodolfo Casás:
Thank you.

Pete Wright:
This has been great.
So once again everybody, thank you so much for downloading and listening to this show. We sure appreciate your time and attention. Swipe over in those show notes for this very episode in your podcast player. We’ve got links to expand your horizons on everything that Rodolfo has been talking about today. And I am sure Rodolfo’s going to be back to talk to me again on this show, and I am here for it. I can’t wait. Thank you again.

Rodolfo Casás:
I can’t wait. Looking forward to it. Thank you very much.

Pete Wright:
It’s going to be great.
On behalf of Rodolfo Casás, I’m Pete Wright, and we’ll catch up with you next time right here on Trilio Insights.