 Hello, everyone. Good morning. My name is Titus Kurek. This is Ryan Bisoner. We're from Canonical. And today we are going to talk about setting up an Edge Cloud in four commands. If you think you're in the wrong session because you subscribed to setting up an Edge Cloud in six commands, you're still in a good place. We just made a decision as we went with this presentation that actually we're going to reduce the number of commands to four because six is too much. So we're going to navigate through the following topics today. We'll talk briefly about the evolution of the Edge computing paradigm. We'll show you the challenges in the field related to implementation of the Edge. And then we'll introduce you briefly to Microstag. This is our new product that we are launching. We'll show you the use cases for Microstag. And hopefully, we'll be able to perform a demonstration where we'll show the Microstag clustering feature. And at the end, we'll basically wrap it up and open up the Q&A section. So before we move on, I'll initiate the Microstag installation process on my local computer. We have a test setup consisting of two laptops to demonstrate the Edge capabilities. So I'm going to start with my local laptop and Microstag installation. While this is getting this going, I thought it would be interesting to reflect on Edge for just a moment. So Edge Cloud is something that means a lot of different things to a lot of different people. And I think it's being defined as end users and vendors come up with solutions to solve issues. And so Microstag is something that we think plays a role in a larger set of tools to solve Edge challenges. When I talk to people about Edge Cloud, some people just need virtual machine management at the pop site. Some people want Kubernetes-style cluster management. Some people want both. Some people want bare metal management. And so again, we think Microstag is one piece of that tool chest to solve the challenges that are common across Edges. And so what we've got going on here is a snap install of Microstag in about seven minutes, I think, it takes to initialize a control plane and a compute-converged cluster. I'll give you the mic back to Titus here. Thank you, Ryan. So while my Microstag is getting initialized, let's talk about a theory for a while. So let's talk briefly about the evolution of the Cloud computing paradigm. So we started at both sides of this picture. We started at the top where we had the cloud, a big thing, usually consisting of tens, hundreds, thousands of interconnected nodes, a cluster, which is typically running in a big data centers with all production-grade features like high availability and things like that. And we started at the bottom where we have the end users, all of those laptops, mobiles, appliances, devices which are usually connected to the cloud. The problem with this architecture was that because of the distance between the end users and the cloud, end users were usually experienced like big latencies, limited bandwidth, and problems like that. So in order to solve these kind of problems, we've started implementing the edge, which is kind of an extension of the cloud down towards the end users. But the edge implementation poses some challenges as well. So this is something that we got used to. We used to deploy a small number of big clouds. But with the edge, we're going to face the following challenge. So instead, this picture on the left, we're going to see the picture on the right. Instead of deploying a small number of big clouds, we're going to deploy a big number of small clouds. So let's say your organization is about to deploy 500 of the edge clouds during a one business year. This basically translates to two clouds per one business day. Do you have tools that are capable of deploying two clouds per day? Do you have manpower to do that? So in order to solve the aforementioned challenge, we designed Microstack, which is a pure upstream, snap-based OpenStack installation. You can get it directly from the Snap Store at snapcraft.io slash microstack. And this is a new product that we are launching at Canonical this week. So nice to meet you, Microstack. But I said it's a snap-based OpenStack installation. So what does it actually mean? So a snap is a universal Linux package that works on 42 Linux distributions. It is basically an application package together with all of its dependencies, with all of its supporting libraries. And it can be easily installed, upgraded, and removed with a single command. And moreover, due to the embedded nature of the snaps, they are also extremely secure, which makes them a perfect candidate for the edge deployments. But the edge is just one of the use cases for Microstack that we see. So where else do we see an adoption potential for Microstack? So we see it in testing. You can install it on your laptop. If you have a powerful laptop with just typical developer resources, like 16 gigabytes of RAM, multi-core processor, some free storage space, you can install Microstack on your laptop by just a single command. So we can do it even right after this demo. One thing you have to remember is it's still fully functional OpenStack. So there is this minimum set of resources that you need to have. But I'm pretty sure it works on my laptop, for example. So I'm pretty sure you're going to be able to run it on your laptop as well. And we also see Microstack in development. So if you are using CICD pipelines, and those CICD pipelines require OpenStack installation during the execution process, then Microstack is a seamless solution to this problem as well. So now moving to the demo part, the OpenStack cluster can be actually set up based on Microstack by just executing four commands, or maybe two, because those are basically the two commands that you execute on the control node, and exactly the same two commands that you execute on the compute node. So this is what I did at the beginning of the presentation. So it's snap install, Microstack classic. This is the command that installs the Microstack snap. I use the channel edge clustering, which provides a clustering feature. Here in Shanghai, I'm installing it based on the pre-downloaded snap because of the networking issues. And then we run microstack.init, which basically configures my OpenStack installation. I have to answer a few questions. Do I want to set up clustering? Yes, I select the role, either control or compute. I set up some password for clustering, and I provide the IP address for the control node. And then on the compute node, the commands are exactly the same. When answering the questions, I set up cluster again. I select the compute role for the compute node. I provide exactly the same password I used for the control node, the IP address of the control node, and I assign the IP address of the compute node. So moving for the demo, let's see how our Microstack installation is behaving at the moment on my laptop. Drum roll, please. So it's finalizing the configuration. As you can see, it configured some of the basic OpenStack services already. It configured Nova. It configured Neutron and configured Glance, setting a seros image. And here we go. So I got Microstack installed and configured on my laptop. Can we have a look at this real quick? Do like a catalog list or a hypervisor list, just to kind of see what this looks like. So this snap bundles in all of the OpenStack clients. Got a little spelling issue there. It does not have spell check. And so you should be able to run all of the common OpenStack clients without having to install anything else that comes along with it. We do have a limited set, I guess, of OpenStack. This is worth mentioning. Nova Neutron Glance sender keystone is pretty much what we've got in here at the moment. It tends to extend that to cover more services. But at the moment, what Titus is doing, he's SSH'd into this other laptop that we've got set up over here. He's doing a snap install on that with the same snap. Once that is done, he'll be able to do a Microstack.init as described a moment ago. And answer the questions to turn that laptop into a compute node profile, basically enlisting it into the existing control plane with two commands and a couple of answers and a trace back. All right. So in that case, it's already pre-downloaded, so we don't have to do the Edge cluster on it. Might I mention that a moment ago, before we came online here, we had a flaky ethernet port. We did that to avoid the Wi-Fi issues. Thanks, demo gods. So here we go. Back to the original machine. Let's see if we can make something happen here. How's the connectivity between these two laptops? This is a prerequisite to have a network. Apparently, that was the demo. So we're at the booth, and we have this working when we have our NICs working. So we'd like to show you more about that if you want to have questions. We are experiencing some networking issues here with these two laptops and the ethernet cable. But we would like to encourage you to come to our booth to see a live demo of MicroStack. We can present it to you. So just with regards to the takeaways from this session, deploying the Edge infrastructure is challenging due to the huge number of independent cloud environments. Instead of seeing a small number of big environments, we're going to see a big number of small environments. And MicroStack is a seamless solution to this problem. It is suitable not only for the Edge, but also for testing and development. We can see a lot of use cases for the potential adoption of MicroStack. This is a snap-based Pure AppStream open-stack installation, which you can test it on your laptop just after this session. So do you have any questions, maybe? Yes? So what's the biggest advantage regarding the DevStack? We know we have a DevStack on the community. I'm sorry. I'm not sure I got the property. So is it about the differences between MicroStack and DevStack? So DevStack is the script-based installation, which is also kind of like a single-note open-stack installation, but it's all based on scripts. MicroStack is a snap-based installation. So it's an application packaged together with all of its dependencies supporting libraries. It's basically an application image that you download, mount locally, and run. That's the core difference between MicroStack and DevStack. Is it working? There we go. So the question was about DevStack and how that fits into MicroStack. I've got a few things to add to that. So in this snap, you have a number of things that go along with it. You're providing a compiled database, Damon, QMU, LibBert, the full suite, ready to go. The other differentiator, I think, is that this is aimed for production use cases. Whereas in DevStack, you're deploying typically from a Git branch, and most folks are not going to be able to lifecycle manage that at the edge over time. So it's not necessarily intended to be a replacement for DevStack, although if you need a control plane up like that, it's a very quick way to get that for a developer purpose. Any other questions? What is the minimal installation? What is the minimal installation with regards to the hardware footprint? So I believe I covered that. But we usually recommend at least 16 gigabytes of RAM, multiple processor, and around 20, 50 gigabytes of a free disk space. You have to remember that it's still fully functional OpenStack installation. My laptop around the corner is an i7 16 gigs with a solid state drive. No. You can install it in a single machine. You don't necessarily have to set up clustering. So you can install it on a single machine. Question. Fast. Well, it does. Yes, indeed. So if you're installing 50 apt packages, you've got decompression that happens and all kinds of things that have to be resolved and installed. So the point is, I think, quite nice. And that is that these are basically pre-shipped in a, well, it's a SquashFs image, essentially, that drops down on. And you start the services. There's no installation of dependencies. It's completely confined, or rather contained, as far as dependencies go. Yes? Ironic. So at the moment, MicroStack supports the basic OpenSack services, like Keystone, Glance, Nova, Neutron. And the adoption of other services is expected to grow as we progress. But at the moment, it only supports the basic OpenSack services. Any other questions? All good. OK. Then thank you very much for your attention. Thank you. And wish you a happy good day.