Multiplayer Infrastructure: How Spacelift and Open Tofu are Redefining Terraform

Hey everyone. Welcome back here to Text Drug tv. You know, uh, we were talking in the Green Room.

My next guest is from Space Lift, and I usually see space Lift at every single cube con we go to. My friend Dmitri Vlado is CMO. I know Dmitri for several companies that he's been with, and we always say, oh, we've gotta cover space Lift, we gotta cover Space lift.

And I never get a chance to talk to anyone from Space Lift. Well, now I finally got a chance. I want to introduce you to, and I'm gonna do my best to get this name right, Mar Walinski Wachinski, co-founder, chief r and d Officer Space Lift.

Martin, say it Maron, say it right for me. Ky You know, it's my, my tongue just doesn't go that way, but I try. Marin, pleasure to have you on here.

As I mentioned, we've been wanting to do more with Space Lift for a long time. Exciting company, exciting story. But before we get into Space Lift, let's hear your exciting story, right?

You're a co-founder here. What, what, what was your passion? What, what drove you to co-found Space Lift?

I will admit that I'm a completely accidental co-founder. I'm a a software engineer by heart and a DevOps person. I spent seven years at Google being an SRE mostly.

And it was like a backend of the backend. It was tape storage, cold storage. So like, imagine it's, you never get talked to unless someone is in real trouble and the only place they can retrieve the data from is your tape that is probably in a vault somewhere.

It is the very, very kind of like not in a spotlight, uh, sort of thing. I did, uh, spend some time in Facebook being building clusters that was more, uh, customer facing, but I'll say as an engineer myself, I was then, um, consulting for our tech companies mainly in Europe around, uh, cloud native, um, infrastructure, moving to cloud, adopting cloud, scaling cloud. And I essentially, you know, I love Terraform, I love HashiCorp Technologies, I love Terraform.

But one thing that was missing is the, the multiplayer mode for Terraform. It felt like it's an amazing product when you're just using it on your laptop. If you're using with a, with with a friend that you can col collaborate on Slack with, and it's probably okay-ish.

The moment you scale to a team and then to a company that falls apart entirely, it's essentially single product. And what we were building with Space Lift is a multiplayer mode for Terraform and other technologies like that. Absolutely.

And, and that was one of the, one of the knock, you know, and, and I'm not knocking, that's a bad word to use. I'm not knocking Terraform. Terraform was, was an, is an amazing tool, but there were things that there were functionality that it needed and, and, but you know, that's a tremendous opportunity, right?

In, in this case, space Lifts opportunity. For those who are not familiar with how this story progresses, terraforms open source. So you wanna build additional functionality, you could build additional functionality and under the open source licensing, you, you can contribute that back to the, to the project and, and, and everyone could use it, right?

Um, then of course the licensing changed a little bit there, and the, the industry, the community kind of, uh, pushed back on the licensing changes. You know, it wasn't truly true open source anymore, and the whole open tofu, right fork came out here and, and that again, gave everyone opportunity to contribute and, and build a, build a better mouse trap, if you will. Martha, and I know Space Lift was one of the key drivers behind the Open Tofu Fork, and you know, of course Open Tofu ISS now part of, uh, CNCF, right?

Its fast growing and you know, I wrote an article a couple weeks back, look, there's probably, there's a place in the market still for Terraform, and there's a space in the market for Open tofu and you know, it's a big market out there. Yes, indeed. I guess my question to you then is though, open Tofu Terraform, what makes Space Lift special?

Where's the special sauce and space lift beyond Teams? Um, so both Open Tofu and Terraform are essentially ized. They're single player, they're meant to be, right?

Run on on your computer, maybe in a ci, but they don't necessarily provide a sort of guardrails. They don't provide a, uh, the coordination collaboration layer. The beauty of Space Lift is that it provides a uniform layer on top of many technologies.

So we do support Terraform, we do support Open Tofu, we support Lummi cloud formation, Kubernetes, you name it. And there is a, there is an, a uniform layer of, for example, policy as code. And that policy as code layer allows you to define your rules and what can go through, what can't go through, under what circumstances.

And not only that, but the policy as code layer also involves coordination. So you can declaratively specify certain processes like how code goes from your, your source control into a particular deployment, right? Where, which way it goes, right?

So this is something that is truly shining at scale, and that's the thing like Space Lift is all about, scale is all about control. It's is all about visibility and auditability, Terraform is a really nice tool and so is is Open Tofu. Um, but they're essentially CLIs and they're not designed to work on a scale where people are using them.

You could build something on top of a generic CICD. So you, you know, we onboard a lot of customers coming from GitHub actions. We onboard a lot of customers from Jenkins.

And to be fair with you, what we're doing at Space Lift is not rocket science, right? Can you build something similar for your company? You can, is it normally your job?

No. You're probably, you know, you're probably having a lot of different things to, to walk on, and you're essentially buying something that is, is good, is proven, is supported, gives you good user experience and gives you good guarantees that it'll stay up, it'll be up, it will, um, be updated, et cetera. There is no magic in what we do at Space Lift.

We're just doing a multiplayer mode very, very, very well. I love it. I think you're being humble though too.

So space lift is, is, it's a platform, right? But, um, you know, and, and so once you have a platform that allows you to interact with projects and products and, you know, one off like this kind of thing, but really, you guys talk a lot about the, the dual approach, if you will. Um, talk to us about what we mean by that.

Uh, is, is that what we launched today? Or you want to talk about intent, right? Or Yeah, well, I, I, I'd like to get into that, yes.

In the whole age agent AI piece of all that, but the twin path is, is kind of it. Yeah. So I think there's, the, the, um, IAC story is, is an interesting one because there isn't really a middle ground in provisioning infrastructure when you think about this.

There are two, like technically it should be a spectrum, but there are two extremes, and we don't have a spectrum. The, the, the extremes are I click on things on a console and it's kind of like, it's fast, but it's absolutely non repeatable. It's not traceable.

You, you have no idea what's happening. Uh, even my colleague will not know I get a database in 30 seconds, but, you know, it's, if someone forgets about it, there it is, right? And the, the opposite end of the spectrum is full ceremony.

Now imagine you're a new engineer on the team and you're asked to provision an RDS database in AWS So now first thing you need to do is maybe you don't know Terraform, you need to learn Terraform, you need to e relearn HCL, but you're not even halfway there. You need to then go to the provider and learn the provider itself, right? Let's say you learn the provider, you know, the APIs provider, and now in, in order to get an RDS database off, you're gonna have like, uh, 10 different resources maybe, and then you need to learn about those 10 different resources.

And then you're not done. You need to write the HCL code. And then for Terraform, you need to set up like a storage backend, like, uh, you know, some S3 bucket to, to store that some CICD pipeline.

You'll plan, you'll review, you'll look at the pr, you'll apply, you'll probably do it a few times because maybe the first time it doesn't work because you're, you're pretty new. Honestly, for someone like a, a, a senior engineer that's a couple of hours of work for a junior engineer or for someone who's maybe senior, but in some other area, like a senior data scientist, that's a few day project just to get that audience database on properly, right? And as I said, there's a spectrum of like, we, we can do it fast and stupid or we can go full ceremony.

And my feeling is that there shouldn't be like just two extremes. There should be options on the spectrum that allow you to do things that are in, in a way that are not stupid, but they don't necessarily involve all that much, all that, that many steps and all that much knowledge that you have to master to even be able to do very simple things properly. And this is where we're, uh, positioning our new product intent.

This is essentially a way of circumventing or short targeting the whole proper workflow. So if we, if, if intent my will to get an RDS databases on the left and that the RDS database existing is on the right and there's the whole set of steps that I need to, to do to get there, then we're, and you know, it, it kind of involves translating your intent into ultimately A-W-S-A-P-I calls so that you provision your database, we're short circuiting that we're taking your intent and we're translating into in, into an a, uh, A-W-S-A-P-I call. And how we do it is essentially we expose an MCP server that proxies open tofu providers.

There, there is a very rich ecosystem that we are co-founders of, of, of all the providers. And now we can short circuit your intent through an LM to an MCP server, directly to an API call. And it feels like magic.

It's extremely fast, and the beauty of it is that it generates the exact same state as Terraform would, minus all of the other steps. Hmm. Fantastic.

Hey, Morrison, I, I gotta do a little housekeeping here. We didn't get into how do people engage with Space Lift? I'm not entirely sure That's the website.

Okay. So best through the website, through GitHub. What, what's the best way to engage with Space Lift?

io. Um, if you're interested in any technical details, you can read our documentation and there's always an option to sign up for a product, uh, free trial. And if you, uh, prefer more handholding, there's always an option to request a demo on the website, and that is probably the best way to engage with us as a company.

Excellent. And as I said before, you were Space Lift was one of the people, but you know, one of the companies, you know, supporting the, the fork with Open Tofu. So talk a little bit about open source.

You know, you, you, you guys support a lot of open source tools, open Tofu, uh, Lummi Ansible, right? These, these are all tools that, you know, have open source components. I talk about spa, you know, and, and then, and then listening to your background.

I'm not surprised, right? Coming Google and Meta and stuff, you know, big open source supporters, how, you know, the relationship between the open source community and Space Lift still? Yeah, so I am a strong believer in Open Source as a good business.

I, it's not just like a, um, an ethical choice. Of course it is an ethical choice, but I do believe that having open standards is a good way of running a business because you, that's taking Terraform into account for open Tofu. Like sometimes people would say, oh, you're supporting Open Tofu, but you are building on, like, you're building a competitor to Open Tofu, or Open Tofu is a competitor to you.

It, it isn't like you can't compare a single player tool to an enterprise platform. They, they're not competing with one another. And in fact, the, the ability of people to use Open Tofu for free turns it into a standard that we benefit from, because then this being a standard, we can support a standard, we can meet community where they are, and by our sponsorship of Open Tofu, we can affect, you know, the, we can touch a lot of that, that, that community.

We get to events. Uh, we are a well-known brand. It gives us the visibility, it gives us, uh, uh, uh, credentials to be able to talk to open Tofu users.

So I, I personally believe that it's a great way of selling products. I know it is difficult, but I don't think any other option is, is much easier. 'cause the alternative would be to try to design your own protocol or try to design your own closed ecosystem.

And then honestly, like people are not idiots. People understand that they're, you're trying to log them in and most of the time they will be extremely cautious and wary of, of stepping into a completely closed ecosystem. The beauty of Space Lift being built on top of Terraform being built on top of PMI or Open Tofu is that you can get in very fast and you can get out very fast.

And that's a good thing, right? Because it, it makes your expectations very clear. It allows you to be kind of courageous with, with adopting Space Lift, it's not a decision for life.

So it is easier to make, it's not a massive undertaking to migrate your stuff to Space Lift because you are already using the same standards that we're building on top of. So you're using Terraform then a migration to, to Space Lift is an, is a lift and shift. And that's exactly what we did with Intent.

Like we are using the community, uh, project, we're using Open tof, we were sponsoring Open Tofu, we are using pmi, so we want to give back and with intent, there are two reasons why you want to do this. First, as I said, standards, standards are good for everyone, otherwise everyone builds their own little closed private Card And private, private thing and, and everyone is unhappy. And standards allow everyone to grow together.

And competition is good. It's actually like, it's lifting all of us, right? Even to a point, like we can copy from our co competitors, they can copy from us.

It's good for Everyone. I agree with you. I, I think what it does, it establishes a floor that everyone could build off of.

And what's nice is it, it's a common foundation, right? And then it, it's up to the individual companies to see where they go. It's kind of like evolution, right?

Uh, Darwin and Evolution and you know, we all start from here, but this one focuses on this, this one focuses on that. And, and that's what makes, you know, a field of a thousand flowers, right? All growing in the same field.

And, and it's an important thing. I want to mention again, you guys will be in Atlanta at CubeCon. Uh, I think it's November.

I forget the date. November 14th or something like that week. Um, look forward to seeing you there.

Aon, thank you for making it on to Tech Drunk tv. This won't be the last time. We'd love to continue hearing more about Space Lift in the weeks and months ahead.

Thank you for having me, Alan. Pleasure. We're gonna take a break here on Text Drunk tv.

Go check out It's Space Lift. Do io you said, right? Correct.

Space Lift io. Go check 'em out. We're gonna take a break.

We'll be back in a little bit here on Text Drunk tv.