Cooper Netease, Rob Richardson, Richard Campbell Carl Franklin discussed on .NET Rocks!
T shirt, visit d d dot dot net. Rocks dot com to get started. And we're back. It's Richard Campbell Carl Franklin run dot net. Rocks to our friend, rob Richardson talking about STO, and this sort of mature model of containers, orchestration dealing with multi clusters and the endless naval terms. They use. It's been so long heaven seen you since an entire commercial ago. Have you been? Seconds. Rob. What even with yourself? We've been shipping things all over the place. Yeah. You could ship quite a lot in thirty seconds. There you go. I love how Cooper Netease has all of these naval themes, and like all of the things building on top of it have now introduced says, as well is the oh, polling in all these things. We're gonna run out of words. We're going gonna have to start using like naval ranks and all sorts of very obscure, naval words that sailors, don't usually say in mixed company, you know. Yes. Naming things is hard. It is hard. Yes. And renaming things is harder. Because it's like okay, do I take the overhead of renaming all of them? Or do I just rename the ones that I bumped into today? What happens about those ones that I didn't bump into do I rename those eventually do? I not now. You can you definitely can get down the rabbit hole with the names. And yeah, it's the only after go so far. I don't know what the heck in this studio is, but it doesn't matter. I can see. I think I can remember this. Yeah. So close enough for the name should we explore? I actually I want to stay on the brownfield story. So we've got a customer, they're running in VM, you're starting to peel stuff out in containers. Do you wait until you've emptied the VM before you light this, upper, can, you do sort of interim phase with a few things running in containers and everything else, still in the VM? What's cool about STO? And I've never practiced it in this way, is that is DO in theory is this mechanism of communicating between things and is not necessarily tied to Cooper Netease. In fact, they very specifically have an implementation tier to get their traffic, shaping methodology into Cooper, Netease, so in theory, you could pointed at some virtual machine services and pointed at some container services. And that sounds interesting of ideally, I kinda wanna wrap all the things in this security blanket. And if it's running in VM, can I ensure that all the traffic coming out of the VM goes through my proxy? I don't know it feels like an interesting exploration. Right. And they I mean they're offering support for different containers services as part of the traffic management. But that's not the real issue. Right. The real issue is getting this stuff that isn't in containers at. All right. I mean in the end it's all Tia, lest HDP. So you should be able to get to it in the end. Yeah it's just data. More data. And I love that in the end is just data, but it's probably running overage to defeat nowadays. So in the end, it's just a web connection says that's kind of like our baseline assembly today. Yeah. You don't you don't really need to go further than that. Should we dive into some of these details about what these entities are within steel, like we know at envoys are, but mixers and pilots and citadels and galleys should we go down that rabbit hole? Yeah. Definitely. We can those start to get kind of weird as we dig out the pieces of another way to go is we can look at. So now that I have this proxy mechanism between my services. What abstractions can I build on top of that? So either path could be really entertaining to dig through. Well, you're the expert so you decide. When I have this mechanism where I've got this control plane, this pilot that drives traffic between containers. Well, now what if I drive some of my traffic to this other container and most of my traffic to the regular.