 Hi, everybody. My name is Sean Madden. I'm an Advanced Solutions Engineer at Piston Cloud Computing. I want to talk to you a little bit about today. I spend a lot of time out in front of customers, giving them pitches about our product, installing OpenStack. And I get a lot of questions about, we've installed OpenStack, and now it's up and running. Now, what am I going to do with it? So I want to talk to you a little bit today about a technology called virtual memory streaming, as well as virtual desktop infrastructure, which is running on top of Piston OpenStack. Let's just briefly talk about VMS in Piston OpenStack and what the benefits of it are. I'll give you a demo of how VMS works, and then we'll move into a solution that we're working with virtual bridges on, which is a VDI solution that is running on top of Piston OpenStack. And then I'll open up to any questions that you may have. So what's virtual memory streaming? So one of the big things people want to do is spin up virtual machines as quickly as they can. The usual process of spinning up a VM is you figure out which instance or which image you want to create your instance from. That instance will get copied in from memory into the local drive. It gets scheduled. It has to get booted. And then eventually it gets spun up, and that takes a lot of time, especially with something like a Windows image. So what virtual memory streaming does, it actually takes an instance that you've created. It creates a live image, which is basically a capture of the state of the CPU, the RAM, and the disk at that time. And we call that a blessed image or a live image. From that live image, we're able to launch many, many instances in seconds, as opposed to minutes, like you'd see from a regular Windows image. And what we do with that memory is when we do boot from that blessed image or that live image, that memory is actually shared across all the VMs you've created. So what you've done is some deduplication, and you're able to approximately double the number of virtual machines that you get on each host. And what happens to when you do this is basically the image is pre-booted. It streams, and you basically have to add an IP address and a host name, and your images are up and running. So it's a very quick process. And for those people that are doing VDI and need a lot of desktops to happen very quickly, this is a very great solution to do that. So some of the benefits of being able to do this with virtual memory streaming, it does eliminate the boot process. This is really important to avoid that boot storm when everybody's trying to boot up their desktops, and it's very high on network traffic. So we definitely want to eliminate that. I mentioned the memory deduplication that happens with VMS. And then you'll see that when I do my demo that these VMs come up in seconds. So it's a very, very fast process to be able to boot VMs very quickly. One thing that we've also done in our product is the VMS is an extension on the KBM hypervisor. And we've also added true live migration with VMS on top of KBM. So we're able to migrate virtual machines from host to host with zero downtime for the end user. So that's a really good feature as well. So let me do a quick demo here of VMS. So what you see here, this is the screen capture that I did of a running cloud. What I'm going to do here first is just boot an instance. So this is just a regular OpenStack instance. And I'm going to spin up. It's a FemRoll. It's a Windows 7 instance that you'll see right now. So I'll boot the instance in. I think I mentioned what's happening is we select the image from Glance. We take that image, and we copy it from our back end to SEP. It gets copied from SEP into the local drive on whatever host. It's going to be a reside on. Once that's done, it has to actually boot. So Windows images tend to take a while. With this screen capture, I sped it up a little bit. And as you'll see here, it took about a minute, there it is, a minute and 43 seconds to actually go from the time I hit Launch to having that virtual machine up and running. So one instance, and this is an M1 small, as you can see on there. So that's a, it's not the smallest type of instance that's in there, but it also isn't the largest. But it took about a minute and 45 seconds to be able to complete that. So what we want to do, though, is if we can, especially if we have a lot of VMs to spin up, we want to speed that process up as fast as we can. So it's done. So it's active. So now what I'm going to do from that image is go, and I'm going to create a live image, or we also call it a blessed image. And what this is doing is now I'm taking that image, and I'm just copying the CPU state, the RAM state, and then the disk state, and saving that. So once that process is finished, now I have an image that's pre-booted and ready to be spun up. So instead of having to do it from the beginning where we have to boot, now all of a sudden that image can start from pre-booted, get its IP address, get its host name, and be up and running very, very quickly. Which, of course, everybody wants to be able to do. So as we'll see here that it's being blessed, now it's active. Now what you'll see is from this live image or blessed image, I'm going to launch 30 VMs. And this will be in real time. So what we saw from the other one is it was a minute and 43 seconds to launch one. This is a three node cluster that we have over at Piston Cloud with a one gigabit per second interface. And it was all hit launch. Once I bring it up, it'll launch. In this one, I think it took about 40 or so seconds to run 30, or to spin up 30 VMs. So now you see they're all getting ready to build. In addition to just spinning this up with VMS, part of the delay that happens with any VM is that it has to go through the authentication with Keystone to make sure that you have the credentials to be able to spin up these VMs. But as you'll see here, I scroll down. The images start coming active when they turn white. So those were originally Microsoft windows images that took a minute and 43 seconds. And now here we are. Some of them are already done and active. The rest are still spinning up. More of them are becoming active. And I think the total time to spin up 30 was about 40 to 45 seconds to do that. So that's a really cool technology called virtual memory streaming. And so why is that important? If we go back in here, I mentioned virtual desktop infrastructure at the beginning. And I mentioned that we're working with virtual bridges on running VDI on top of Piston OpenStack. And this is actually really cool. So what I have here are some screenshots of the interface of the BridgePoint solution from virtual bridges that's running on Piston OpenStack. So what we see here is we're able to create a cloud from this interface that we're going to be interfacing with for VDI. So we're looking at the clouds here. And with this solution, you can actually see multiple clouds. So in this case, we have a Piston OpenStack cloud. There's also a public cloud shown here. You can have multiple clouds and be able to interact with all of those clouds through this BridgePoint interface. And again, with software images, so these are the images that are starting glance on the Piston cloud that are being shown here. And again, even with the images, you'll have images from multiple clouds that you'll see under these images. And these images can run on different clouds. So these could be the Piston OpenStack images. These could be public cloud images. These could be any number of images from many, many clouds. And once we do, going back to VMS, once you create the first blessed image and the pre-booted image ready to go, that can be one of these instances that's here in the VDI solution to be able to spin up. So you can spin up Windows images, Linux images, very, very quickly through this VDI solution. And then so the virtual desktops, they could be created by one of the administrators. So this is just a screenshot of the page and one of the desktops that has been created, a virtual aid desktop to be spun up when you need to get it up and running. And finally, this last picture is just a picture of the quota summary. This will show you for all the tenants within these different clouds, whether it's the Piston cloud or the public cloud or any number of clouds, what your usage is, and how many more desktops you're able to spin up. So the whole point here is, number one, a lot of people ask me about OpenStack, it's up and running, how do I do different things? VDI is a solution that a lot of people are asking about. They want to spin up multiple desktops within their companies and move everything from being physical desktops to virtual desktops. VMS is a great solution to do that because it definitely helps speed up the process of booting these images from the old fashioned way of just booting them up from scratch to instead building these live images that are pre-booted and spinning them up very quickly. And it helps to have a solution like what virtual bridges has put together with VDI to be able to access into Piston cloud or any other cloud to be able to spin up these images and have access from one user interface out to all of these clouds. So that's kind of what I wanted to give you guys a intro to. And if you have any questions, feel free to ask. And if not, I thank you for your time and I'll stay for a few more minutes. Thanks a lot.