 Can you guys hear me? Hello? You listen? OK, that's working. So welcome. Thank you, everybody, for coming. So my name is Rafael Rosa. I work for the internet app. And today I'm going to talk about our cloud product, which is using OpenStack. So just a sec. OK, one more time. Sorry. Wrongs I said. So I work for internet app and iWeb. We are a hosting company, an internet service company. We provide infrastructure services like co-location, hosting, and cloud and network, network services like CDN, IP, high performance IP, and other things like that. So we are hosting more than 13,000 clients on data centers spread around the world. And we have data centers in North America, in Europe, in Asia, and we have more than 750 employees. In December 2013, internet bot iWeb, which is a Canadian company. And that's where I'm from. And so that's it. I'll show you briefly what do we provide as products. So our portfolio is quite big, and it's focused on two main things, performance and flexibility. We provide all those products, like cloud, bare metal, dedicated server, and co-location, and allow our clients to use them in conjunction. So you can have, for example, bare metal server and co-location systems and make them work together. And the same goes for cloud. All of this is provided with accelerated networking that we have from our Pops management services in a single control panel that you can use to manage everything. So these are the places where we have data centers. The ones in green are the ones that are managed by internet, owned and operated by internet. The ones in blue are the ones that we have partners around the world, and we are always expanding our network. So we're mostly concentrated in North America, but we are expanding. But how did we get here? How did we get to the point that we're talking about OpenStack? So in the beginning, the internet and our web were pretty much a traditional hosting company. So we offered hosting, co-location, network services. That's what everybody does. That's very traditional. And then came a time when the cloud became a thing, the future, the future of hosting is the cloud. And we decided to take a leap and provide a new product that is going to take advantage of all the cloud things like self-serve purchases, sorry, API, paper use, and other things. So to do that, we chose to use OpenStack. OpenStack is a good platform for what we want to do. It's created by the brightest people that we have, a big community, a welcoming community, a community that wants to share knowledge. And that's very good for us. We don't need to reinvent the wheel every time that we want to create a new system or we want to provide a new service. We can use the expertise of this community to help us provide a better product for our clients. And back in 2010, that's when we started looking at OpenStack. So we already had some commits done in Baxar. So since Baxar, we are already contributing code. And we are trying to work with the community to provide a better OpenStack for everybody and trying to improve the ecosystem for everybody. So here is a brief summary of all the contributions that we made. We had a lot of contributions, also during the Havana Grizzly period. And during this time, we had lots of contributions for Cinder, for example. We use SolidFire, and we provide a lot of help to create the Cinder driver for SolidFire. And we committed code for that. And we also helped a lot with the Stack Forge project, which is the puppet automation for installation of OpenStacks. One of our contributors, by the way, is Macheganier. Perhaps you guys have seen him during the videos that were shown today during the keynotes. He was elected super user because of his contributions, both on code and also as a reviewer. So we are very proud of the contributions that internet and I web made to the OpenStack project. So OK, this is how we got to use OpenStack. This is how we interact with the community. But how do we make OpenStack something that is unique, that is different from what we have on the market, that's better for our clients. And it provides the values that we see on the internet. We decided to get OpenStack and change a little bit how we deploy it. We chose the right stack for the job. In this case, we decided to use an all SSD storage solution because we want to have the best performance for our clients. So we offer two different kinds of flavors. You can have ephemeral storage with instance SSDs, or you can have persistent storage using solid fire. On both situations, you're using SSDs. And we're providing the best performance that we can on the storage side. The other thing that is very important for our clients is performance guarantees. Our clients come to us because they want the best performance. They don't want to suffer from issues like noisy neighbors or paying for performance that they're not getting. So how do we do that? We created different flavors that provide guaranteed CPU threads for each one of those clients. And we don't oversell the CPU for those flavors. And that allows the clients to use full thread of CPU when they choose those flavors. On the storage side, we provide iOS, QOS, on IOPS, on solid fire. For each volume that you create, you can set up how many IOPS that specific volume can use. And that way, we guarantee that the client will have access to the performance in the storage IO that he needs when the time comes. Another thing that's coming soon is that we're going to provide, through the same API and through the same control panel, bare metal flavors running on OpenStack. So instead of creating a VM, you'll be able to instantiate a bare metal machine that will be provisioned in a couple minutes. And you have the full VM for you, no sharing. You're the single tenant inside that VM. And for some clients, that's very important because you can extract all the performance that you need from the single machine. Another thing that we did, and it's very important for us, is how can we deliver the best networking experience for our clients? We do that by improving the network and creating cross-platform setups. So what happens is, as I mentioned, you can have machines running on our OpenStack cloud. And then you can have other different servers, like database servers, running on bare metal machines on the hosting side. Or you can even have legacy infrastructure running on colocation. And you want all those things connected. You can do that through our networks. You create those machines, connect to them, and you have a private network that you can use to transfer data across all those infrastructures. This allows you to choose the right infrastructure for each job. And we also use our mirror technology that provides the best. It provides the low latency connection by choosing which ISPs it should use for each client on a real-time basis. On top of all this, you have a single pane of glass that is a single one place that you can manage all those resources with the same features, with the same UI, and with the same support. And that's very easy for users because they don't need to learn two different tools to use our infrastructure services. So in the end, what we were able to create is an evolution. We can allow clients to move from the cloud to colocation and vice versa, depending on their needs. So let's say that you have small deployments, or you have elastic workloads, or you create a new infrastructure. All of these is much easier to do if you're using the cloud. You don't need to think about it. It's very easy. Click, click, click. In five clicks, you'll have it working. It's a little bit more complicated if you need bare metal. But we allow them to evolve. After some time, perhaps, you need more performance, or you need dedicated servers. We allow you to add those servers as you need and take advantage of them and go up to colocation. That's very good for people moving legacy workloads that are running on old machines, but they want to use the cloud. And they don't need to choose one or another. They can mix them as they need. I'll use a couple of clients that we have using our OpenStack platform to showcase how flexible it is. So the first one is Carl Jr. This is a very interesting case. He's a 17-year-old guy that is a pro gamer. He is the world champion of Trackmania 2 Stagen on the Electronic Sports World Cup. So this is a 17-year-old kid that used the cloud to create servers so he can train with his friends and alone. This guy trains six hours, eight hours, doing this racing game every day when he's training for competitions. And he uses the cloud. He's a very small user. One cloud is good enough for him. We allow him to do that on our cloud. That's the card that he uses sometimes when he's playing on games that are sponsored by iWeb. Another use case, which is much bigger and very interesting is Cortica. Cortica provides real-time image recognition software that can be used, for example, to recognize a picture and choose the best ad that goes with that picture. Imagine that your clients are uploading a picture of a beach. So their software can recognize that and then provide, for example, the best travel ads based on the picture without anybody needing to add meta tags to that picture. And that's very cool. It's automated. Everything's automated, and it uses a lot of resources. They use machine learning to accomplish that. And most of the times, they use bare metal machines to do this because that's the best cost benefit for their use case. But the thing is, they also use the cloud for POCs and dev and test. So we give them the flexibility of using the cloud to access their bare metal machines and make their systems work. And they don't need to choose. They are choosing the best tool for the job. The last case is Boink. Boink is a company that provides unlimited music streaming and downloads for mobile apps in Venezuela and Chile. So they provide this limited streaming, and they are hosted with us. And they use all our products. They use a bit of cloud. They use a bit of hosting. They use a bit of colocation. And they use our networking services as CDN. So they use CDN to deliver the media to their clients. They use the cloud for some tests and dev and for some elastic workloads that they can spin up and down when they need. They use hosting and hosting for their media, for their media storage. And they use colocation as DR and also other types of media storage. So they use all the services together and they can deliver the best solution for their clients. They can deliver the best performance for their clients and they can use the right tool for the job. It's always the right tool for the job. Our main focus is delivering performance and flexibility, right? So today we announced the next Generation Agile cloud that is GA today. It's running on OpenStack Havana. You can have access for our site. You can use the API to provision new machines. Just go up and sign up. As we told you, it's all SSD-based. The network is designed to work in hybrid environments. We will have bare metal flavors coming soon and you have the single pane of glass to manage all those services. Be them cloud products, bare metal or colocation. So that's what I had. Thank you very much for coming. If you want more information, please visit us at booth D7. We'll be there to provide more info and to clarify any questions. Thank you.