Tech Lounge
00:04:49 · Will Stewart, Chris Chinchilla
[00:04:49] S1: today or tomorrow we'll be announcing that we've recently hit general availability of our Civo integration. So with Northflank, you can start with Civo, you connect your Civo credentials, and then we can do the full cluster lifecycle management of your Civo cluster. Boots up in about ninety seconds, then Northflank is installed on top. And in under ten minutes, you have a production ready workload platform where you can deploy databases, highly available services, and you can just get back to doing what you do best, is building your application. [00:05:16] S0: Know any company founder hates this kind of question, but even in the vendor hall here, there are some companies doing similar things, some of whom I actually know way better. So again, this sort of platform as a service, platform team replaced your DevOps engineers, etc. There's still a few companies doing this and you've given some details, but in comparison to those who also say we focus on the developer experience, etc. They all say that. So really, I actually can see some things on your more technical details that I think are more interesting. But where would you really say your points of differentiation?
00:05:59 · Will Stewart, Chris Chinchilla
[00:05:59] S1: I think the proof's in the pudding. We've got hundreds of API endpoints. Every single feature we've added to the Northflank platform is available in a real time beautiful UI, an API, CLI, and GitOps. So if you're a platform engineer or you're a platform leader and you want to build something special, you want to build a platform for your team, you just leverage Northflank's APIs. Or you can have those engineers then self-service to a real time API. So we've sort of got like a platform API that some of our largest customers and enterprise customers are using today, and they looked at the market and saw that Northflank was the only team offering that level of abstraction, this higher level of abstraction of workloads, but also giving them the capability of some of these sort of platform I'd say if we look to the past, people like Cloud Foundry, Pivotal Cloud Foundry, and Heroku, very successful platforms. Cloud Foundry, sort of defeated by Kubernetes, and then Heroku sort of sold too early to... [00:06:55] S0: To defeat themselves.
00:06:55 · Chris Chinchilla, Will Stewart
[00:06:55] S0: To defeat themselves. [00:06:56] S1: Exactly. Then stopped innovating. So I'd say that rather than we've come at a different angle, some of the fine folks in the room are building, and some don't necessarily have to name some, but very effective at marketing, very effective at getting the word out. And then we say, Well, where's the product? Where can I actually use the product? And I think that's where Northflank, we've come at it from a different angle. We spent five years building a very sophisticated, very well thought product that's rich in-depth and breadth. Self-service gets started in minutes, and then we have teams deploying a thousand microservices on the platform, but we also have mothers deploying glucose monitors for their kids. So we've managed to solve the developer experience where you can self serve and deploy complex applications with zero dev experience, all the way to engineering teams of 50 platform engineers deploying very complex workloads. [00:07:46] S0: Yeah, that's an interesting one. Let's come back to that in a second. Just one final question on comparing to the competition. Often in places in products like this, the biggest competition is actually people just doing it themselves. We kind of know there is a business justification for not having your team doing it themselves. But often that comes later. Yep. So yeah, you know, why not just use Kubernetes yourself or Open Tofu or Pulumi or, you know, a whole bunch of other options, you know. Why not do it yourself?
00:07:46 · Chris Chinchilla, Will Stewart
[00:07:46] S0: Yeah, that's an interesting one. Let's come back to that in a second. Just one final question on comparing to the competition. Often in places in products like this, the biggest competition is actually people just doing it themselves. We kind of know there is a business justification for not having your team doing it themselves. But often that comes later. Yep. So yeah, you know, why not just use Kubernetes yourself or Open Tofu or Pulumi or, you know, a whole bunch of other options, you know. Why not do it yourself? [00:08:18] S1: So I 100% agree. 99% of our competition is DIY. That's even what we tell the investors. It's the truth, There is no breakaway product yet, but that's also a good thing for us. That's an opportunity. People are doing DIY because that's credibly the only option for them. You can't buy a product that doesn't it has to do everything. If you buy a product and it doesn't do 15 of your 100 tasks you need it to do, that's not good enough. And that's the methodology when we think about product development at Northflank is to work with customers very closely to understand what do you actually need. Do we actually meet your objectives? If not, let's rapidly build those features.
00:08:18 · Will Stewart, Chris Chinchilla
[00:08:18] S1: So I 100% agree. 99% of our competition is DIY. That's even what we tell the investors. It's the truth, There is no breakaway product yet, but that's also a good thing for us. That's an opportunity. People are doing DIY because that's credibly the only option for them. You can't buy a product that doesn't it has to do everything. If you buy a product and it doesn't do 15 of your 100 tasks you need it to do, that's not good enough. And that's the methodology when we think about product development at Northflank is to work with customers very closely to understand what do you actually need. Do we actually meet your objectives? If not, let's rapidly build those features. [00:09:01] S0: Okay. Alright, let's dig into a bit more technical detail because I think there's some interesting things. You already mentioned... I'm gonna get the list, but you already mentioned one of them. Yeah. You mentioned that you use catted containers by default. Yep. Now, I don't know if other companies do and just don't mention it, but you explicitly mentioned it. And I thought that was an interesting place to start with some of your platform toolkit options because you've got some interesting ones here. Why did you decide to go with Keta containers over? I don't know, what to call the alternative, but... You guys have normal containers. Yeah.
00:09:33 · Will Stewart, Chris Chinchilla
[00:09:33] S1: One of the core principles of building a multi tenant platform is security. Essentially, platform like Northflank, a platform as a service, whatever you want to call it, is running untrusted user code. Yep. You can't have untrusted user code just masquerading on top of Kubernetes clusters. That's a recipe for disaster. You know, customer a, attack customer b, or noisy neighbor, whatever, it's a recipe for disaster. So with Northflank, from the get go, as soon as we released production, we had a secure runtime. At the time we were using gVisor, a tool maintained and built by Google. It's a fantastic it's wonderful. And then at one point, you know, we were seeing performance impact, we were seeing sort of load on customer websites where they weren't happy with performance, and we sort of evaluated what was on offer. And Carta containers is a sort of abstraction of KVM on top of Kubernetes. It's a fantastic platform. And we've been live in production with Carter, deploying millions of pods now for a couple of years, and it's very performant. And then for some of our customers who need to run on AWS, AWS doesn't let you run nested KVM nested virtualization, so we have to support gVisor for sort of on demand AWS. Was wondering if [00:10:44] S0: there were any I haven't had any massive experience with it, I wonder if there were any compatibility issues. But I'm guessing if there are, it's probably showing it's a bad sign in some cases.