Hybrid-Infrastructure from the Ground - Maciej Lelusz

Machine-generated from the episode audio and not hand-corrected, so names and technical terms may be imperfect. The audio is authoritative.

Hey, you're listening to another great episode of the CTO Advisor, where we're coming out of the quarter four of 2021. I'm coming off of AWS reInvent. With me is Magic. How's it going? Everything fine. Okay. Thank you for the invitation. So pronounce your full name for me and your company. My name is Magic Lelouch. I'm CEO of Inleo. Lelouch. I love the last name. I actually took a semester of French in high school, so I at least understand the accent.

I don't know if I can actually say many words in French anymore, but you know, there's that. If you haven't gotten that so far, Magic is based out of Europe, a lot of European customers, and I met him via a mutual acquaintance, someone we had on the podcast, Chris Williams during AWS reInvent. Are you an AWS hero? Yeah. Happily, yeah. It's the old hero. Yeah. Very good. So you're an AWS hero on, and why I wanted to have Magic on was to talk about the reality of hybrid cloud or hybrid infrastructure.

We'll just call it hybrid. From the field, you run an organization full of architects who deal with both VMware, classic VMware environments, and AWS, correct? Yeah, exactly. Exactly this year. What are some of the typical challenges your customers will come to you with? I think the most important is that they want to consume new software stack. They want to go for cloud-native applications, but they still have a lot of business inside their old legacy, like now everybody's saying like this, like, oh, legacy, creepy data centers, but basically, it's giving them a lot of juice.

They cannot cut this technological background that they're coming from, and they just go for a cloud-native because it's too expensive, too risky, but they see that there is a value because the new people who are coming to the IT, they want to work with those technologies. They go for hybrid to try everything that is good from both worlds. They have, for example, this VMware-based stacks solutions, wherever vSAN, big vSAN clusters and vRealize automation farms, and from the second side, they're going for VMC a little bit.

They're going sometimes mostly for cloud-native solutions, for example, like EKS or serverless solutions because they see the value there. Now, the idea is to get those two environments to interoperate, so the site one is traditional VMware stack that you just described. Site two is AWS, a quote-unquote site or a set of AWS VPCs, and then there's some type of direct connect in between? Yeah. I think so. The challenge is networking, but not anymore, I think. The world push a little bit boundary there, and we have a lot of entry points to the public cloud.

There is a lot of knowledge on the data centers part of the business that we are cooperating with, so there are a lot of super experts regarding the networking. They are still working. They are still alive. It's not something that's like, they are here, we are here, and they are able to design those networks, like BGP peering and all of this stuff important to make the packets flow from the cloud to your data center and back. The networking seems to be something that is challenging, but you can deal with that, and then you can consume the new infrastructure, and it's not so far away in the end.

What you just described is, I think, one of the core pieces of, not necessarily data, but an observation that I've noticed over the past year or two, that if you're born in a cloud, you kind of don't think about any of this stuff. It's a box. I put stuff in, take stuff out. It's a box. Whereas in a hybrid scenario, if I'm moving from traditional data center to wanting to consume AWS, not so much. I have the expertise. I have the chops to do the networking stuff.

When AWS says that I have to set up BGP, that's not really that scary because it can get complex, but it's not as complex as my data center networks or my multiples of data center networks. They've kind of abstracted a lot of the complexity from me. What have you noticed in the difference in culture between companies that are born in the cloud and companies that are kind of migrating or adopting cloud technologies? No, I don't know if I use the proper word, but I think that companies born in cloud are ruthless.

They just consume it. They just take everything what they give them and it's amazing how they grow. But sometimes when it's about the governance, about the security, those stuff, not maybe from the very beginning, they are there because you can use so many stuff and it's so easy that you can explode with your idea to like a thousand, millions of customers very rapidly. And then later on, you know, that small security is going there, you know, that the governance, you saw the re-event this year.

I think that they stop to talk about security, they start to talk about governance. There was this, there was the conversations there, how to be sure that what we are doing is right because we already did it, you know? And in the data center world, it's not so easy because you need to think a little bit in forward. I'm not saying that it's better. I think that it's just different, you know, that you need to prepare everything because your resources are limited.

Your technology is more or less complicated and you need to have people, governance, sometimes even before the project starting. And you know, that's, I think that's the difference and it's a big difference, the big different mindset. Yeah. I wrote, not wrote, I read a really great blog post from a former AWS engineer who was only an engineer for 10 months, an application developer. I think he was actually a design interface developer. He was only there for 10 months and kind of the gist of the blog post or the title was about his short stint at AWS, but what it really exposed was the culture of AWS when it comes to development, planning, governance, et cetera.

So as you just said that I put two and two together, it's really funny because AWS's culture is very much like the culture of a startup where there's not necessarily a lot of planning upfront. There's a problem. There's a technical solution to the problem. Let's execute the technical solution to the problem. Whereas if you're older IT shop, you know, you've gotten, I don't want to say burned by that. You've learned the lessons from it over the years that that is not, it's not sustainable.

I think is the, probably the best way to put it. I don't know how AWS has figured out how to sustain it to 270 plus services, you know, for 12, 13 year period, et cetera, but they've, you know, they, they worked it out. So let's shift back over to kind of the, the technologies for, can you just oppose or tell the difference between the technologies in the stacks or the operational differences? I want to focus on the operational differences between a born in the cloud company and a company that's more hybrid.

I think when, when you go, when you, when you are on the beginning of the journey to set up a hybrid infrastructure this is the thing that you, yeah, that it's the people talk a lot about the day two operations because they are really complicated. It's not in the sense that it's impossible, but you need to sustain sometimes two different teams that they have two different competencies, you know, and for example, when you talk about AWS and the native solutions, different mindset, but also different knowledge base, you cannot take those people, you know, and operate inside the, I don't know, vSAN, hyper convergence platforms to different worlds, different tools, different, different mindset again.

So the day two is challenging, but it's all about what you want, what the advantage, this set up for a hybrid cloud gives you. If it gives you ability to give to your different teams, the products and solutions that they're really eager to use, and it's perfect, you know, they will manage to do that. If they work in DevOps, let's say methodology, let's call it like this. They will figure it out. And when they will have good leaders and the old world, you know, like legacy data center world of, you know, the VMs that they will stay here, well, a couple more years for sure.

They still will have tools to at least have this elasticity of the cloud. So they will be able to move somewhere because come on, you cannot buy server right now, at least in Europe. It's taking you like 300 days. I'm trying to buy a server right now. I cannot physically buy a server for a couple of months. And when you are a big enterprise, you need to take this risk into the consideration. So the cloud is solution for you for at least that risk.

It's many more, but you don't want to take your entire, you know, technologies like Android to the trash. It costs you too much in over those years. So all those people wants to work with the tools that they used to. And what I realized on the last event and to the conversations, that's the very challenging part to create tools that they are understandable for in both languages. And they'll solve the same, you know, like they do the same job. Then what seems to be even a hybrid of that.

So, you know, Gartner would have called that bimodal IT a few years ago. I don't know what they're calling it now, but I think that's a fairly reasonable assessment of where we're at. But adjacent to that seems to be this concept of cloud native in the sense of Kubernetes. So, you know, there's a push to take your VMware stack both on Kubernetes and then give that cloud native, that consumption based experience for developers for that. And there's also the desire to take Kubernetes and bolt it on top of public cloud and abstract the public cloud and kind of get the best of both worlds.

So if you're thinking hybrid, oh, I just have Kubernetes and I can give Kubernetes to my developers and life is good. I don't know if I've experienced a shop where that's actual reality. I think there's some complexity in that. Have you run into kind of this architectural desire to paint over the complexity of hybrid cloud with a Kubernetes? You know, I always during that conversation showing the Pollock paintings, you know, as an architectural reference, because it's appealing to the business to have, you know, one layer to do one ring to rule them all, you know, and I just dropped the developers there.

I dropped the developers there. But in the real world, you know, all this little lines, you know, they go, they look like a spaghetti in the end, you know. So from an architectural point of view, it's it's challenging. It's a lot of networking part there, which sometimes it's not yet figured out. Sometimes you need to fight with it a lot with from the infrastructure perspective. But I think the more work to do and a lot of my colleagues doing that part is to like have a relation with developers.

I think like DevRel right now that you need to just change the way that the code is made and it's going to work here. It's going to work there. If they will, you know, like in the developer world, the developer will design the code the way that it's easy to be moved. You know, it's not associated to IP address, not associated to the things that even we are we that we are coming out from the legacy world don't like. We can say that we used to that, but I like it.

I don't know if I like it. You know, I just I just used to it. And if I if I look at, you know, AWS stack, for example, in the cloud, how it looks, how it's how it's operate, you know, I like I love it. You know, I love the way how they do it. And to move it to your own data center, it's, let's say, almost impossible from the architectural point of view. The solution is not there, not here.

The knowledge, it's only the big, you know, cloud vendors like AWS or Azure or whatever. And because of that, when you want to build something very similar in your own data center, you will hit challenges. And sometimes the challenge is that there are not consumer technology here that you can just take from the shelf, from the box, you know, and using your data center. It's not existing. But writing it, you know, yeah, I've I've run into that awful lot that the big picture and it's not just Kubernetes, it's open source in general.

And I'm not going to pick on all open source. I think it's infrastructure, open source products specifically are designed for. Not even design, they lend themselves to big shops, shops that are able to hire talent to put additional value on top of existing products. There's rarely, you know, kind of the open shifts of the world, which takes a big project like Kubernetes and provides an opinionated platform for you to deploy a VMware Tanzu. But even those are lacking in features. When you look at big cloud providers, AWS is so ahead of the curve.

You know, when you think about, you know, SQS or Aurora or some of the other features that they manage and they add value to and getting that out of your open source project like a program on premises, it's, you know, it's not as you said, I can't go I can't go to VMware and buy Aurora, but it doesn't exist. The thing is, if I may, you know, like all of these things that we talk about are tools. The biggest challenge for an architect, for a guy like me, is to understand the business, what they really need, because sometimes business don't really know what they need, you know, and to define that part, it's just showing the tools are out there.

And you just need to take them, you know, from the shelf. If you want Aurora, don't build your own, you know, you just will burn so much calories. And okay, you can, you have challenge to collect with that. But like you said, okay, some BGP sauce, you know, some sparking of IP address, whatever, you know, it's not something that is like a crazy heart. It's just that the issue which I'm facing very often, it's like, let's try to build something that is already existing.

And it's awesome. I'm saying this is just tools, you want a public cloud solutions, use public cloud, you have business, you know, justification to use legacy data center, I don't know, maybe you're, you know, like an army or some special force that you want to really have those data together. Not perfect, but just don't stop yourself and to use the public cloud or use only data centers. That's about the mindset. What I want to come back for a moment. Technology is not religion.

It's changing all the time. And if you will keep stay where you are, I'm the expert, you know, like 12 years already. So I'm really connected to VMware, but I'm not saying that VMware is perfect for everything like AWS is not perfect for everything. But there are very good tools out there that you can use to make your business better. Just not be religious too much. Yeah, I think that's a whole nother podcast episode, community versus religion. I think one of the byproducts of, it's not just that I can go back to my Novell days.

It's not just the V community or the data center community. Those communities are tight communities. And there's kind of a sense of betrayal when, you know, somebody selects a technology that's not part of your, you know, community that you identify with. It's not, this is not a religion. It's just, you know, it's a tool like I don't, I don't, I don't think I, I don't think the Bosch hammer versus the craftsman hammer really matters. Which hammer I use to nail the nail in, just as long as I nail the nail in, I built the thing that I'm trying to build.

Of course there is a brand that you can love and you can go for a brand and you can buy like a silver hammer or golden hammer, wherever, but it's still gotta be a hammer. The most important thing is that's what you like, you know, and it's not that you really believe. And I think that's, that's the point when you, when, when you talk about hybrid cloud, don't close yourself in the silo of one trend or only containers or only VMs. Come on, everything is open.

These give us plenty options to go for. So Magic was really great catching up with you. I really appreciate you spending your late afternoon with me. It's early morning here in the US, late afternoon in Europe. For those who want to reach out to you and your company, how can they find you? Yeah, I think Twitter is the best way, like everybody. So Inleo is at the Twitter. I will send the links. You can publish them. It's going to be perfect.

And I'm also running my own Twitter account, so you can find me there definitely. com, that's the place where you can find our website. Yeah. Thank you Keith for that. All right. And thanks again for listening to another episode of the CTO advisor. You want to learn more about the CTO advisor? com. My DMs are open. If there was a question around this podcast or hybrid cloud in general, you know, we run the CTO advisor hybrid infrastructure.

We go on this journey that customers are going on trying to integrate traditional VMware environments with the public cloud. DM me at CTO advisor on Twitter. Talk to you next CTO advisor podcast.