 All right, so we're going to go ahead and get started. How's everybody doing? Still awake? All right. So my name's James Lobaki. I work in the Cloud Business Unit, Red Hat. And this is Olai Gona. Yeah, I'm the co-founder of ManageIQ, and I'm now running software engineering for Cloud Forms at Red Hat. So today, we want to talk to you a little bit about unifying the management of OpenStack with public cloud providers, your existing data center virtualization, and some other concepts as well. So for the agenda, we'll take you briefly through something called Cloud Forms that Red Hat has, and what we mean by hybrid cloud management at Red Hat, and where it fits into Red Hat's OpenStack management strategy. And then we'll dive into some of the specific features Olai will take you through the platform architecture, what it is. Then we'll go through a couple of demonstrations of the OpenStack support that we just recently announced. All right, so Red Hat recognizes that the role of IT is changing, right? IT is becoming a service provider to the business, so no longer are we serving up machines on physical infrastructure to end users. They're actually expecting IT to basically deliver these services through private cloud, public cloud, and even IT using public clouds to host services. So really, it's this evolution to hybrid IT that we all recognize. So again, while the IT consumer used to be really consuming applications and services on traditional infrastructure, physical and virtualized highly, really what's been happening is we all know the IT consumer has been using public cloud and public infrastructure as a service, right? That's what this has done as the IT consumer has been using public infrastructure as a service and public pass is that their expectations are beginning to rise. So that's creating a huge demand on enterprise IT organizations to deliver self-service, elastic infrastructure, rapid environment delivery, and accelerating service delivery. So all this pressure is starting to come on to IT organizations to do this, right? And at the same time, IT organizations are trying to balance agility and openness and choice and leverage their existing investments, which is why you guys are here, right? So understandably, when you start talking about OpenStack, it's IT organizations and service providers as well trying to build more efficient IT infrastructures that can deliver these capabilities of self-service and elastic infrastructure. So and underlying that, they're interested in building shared services between their existing investments to bring those to bear. But what we're going to be talking about today is really this hybrid cloud management layer. So while you don't really solve a fundamental problem of complexity, if all you do is add in another silo of efficiency with something like a private cloud technology, likewise with platform as a service. So what we're really going to focus on is the hybrid cloud management layer. So why is this important to the folks at OpenStack, right? You might be thinking, how is this relevant to you? Well, one is if you don't have a hybrid cloud management platform, it's very difficult to migrate existing workloads and determine which workloads are best to move on to a private cloud that's based on OpenStack. And likewise, if you're a public cloud provider that's basing your public cloud on OpenStack, understanding having a hybrid cloud management strategy that allows your users to bring their workloads to OpenStack is important. So when we talk about OpenStack or Red Hat, there's really kind of three key areas that Red Hat works on OpenStack. First is through Red Hat Enterprise Linux OpenStack platform. So that's providing a stable, secure lifecycle for OpenStack. So if you're familiar with Fedora to Red Hat Enterprise Linux, you've probably heard Brian and Mark talking this morning about how we're basically doing the same thing with Red Hat Enterprise Linux OpenStack platform. The second is through instrumentation, APIs, and allowing for automations. This is all of our work we're doing in the community itself. So again, you heard Mark talking about Tusker this morning in triple O. And there's a number of communities, including Solometer. This is really all the instrumentation that's going to be required for us to be able to bring a cloud management platform on top of there. And the third area we're working is the cloud management platforms. This is Red Hat CloudForms, which then basically leverages the instrumentation, APIs, and automation to provide a hybrid cloud management solution on top of OpenStack. So the capabilities that our hybrid cloud management platform delivers are all these. So approval and workflow, compliance, self-service, basically all the operational capabilities you'd expect. And with regards to OpenStack, we're doing this by integrating with Nova, Solometer, AMQP, Heat, Glance, and Puppet are actually future integration points. So today, we're integrating with the Nova API, Solometer, and the AMQP event bus. One thing to note is that CloudForms actually supports both your data center virtualization, so VMware vSphere, Red Hat Enterprise virtualization, as well as public cloud on Amazon EC2 as well. All right. So handed over to Oleg, he's going to take you guys through the platform architecture here. So the CloudForms product is built on an acquisition that Red Hat made last year of ManageIQ. The platform architecture is kind of a revised LAMP stack. It's Red Hat Enterprise Linux, Postgres, Apache, and Ruby on Rails. We deliver CloudForms as a virtual appliance, a virtual machine. So typically, people downloaded it, and they're installed minutes later. A lot of our customers, when we go on site and do POCs, a lot of them are seeing value within the first hour, the first two hours. In terms of scalability, I'm just going to give you guys an overview. So there's a virtual appliance. Inside the virtual appliance, you can have workers, workers to process within a virtual appliance. There's many different types of workers, probably 15 or 20 different types of workers, each with its own kind of role of what it does. Each appliance can have a role. And based on that role, some workers can will be started or won't be started. You can also designate some appliance roles to be primary on one appliance, and secondary on another appliance so if one appliance goes down, it'll fail over to the secondary. There's also a notion of zones within CloudForms, which allows you to map providers to appliances. So if you have, say, one open stack instance and you want three appliances to point at it, to do different things, you can configure that to happen. Or if you have a vSphere or whatever. And finally, there is a notion of a region which holds the data for all the zones. And you can actually do multi-regioning with replication going up to a master region for reporting on many different installations of CloudForms. So one of the things is that basically, the way we've kind of packaged this up is that we deliver everything as something called Red Hat Cloud Infrastructure. So you can go pull the hybrid cloud management yourself. But there's also an offering where we basically provide either Red Hat Enterprise virtualization for your data center vert, along with Red Hat Enterprise Linux open stack platform. So if you have cloud-based workloads that you want to use, with the hybrid cloud management as well. So just to take you through how this would typically be deployed, so you guys are probably familiar with data center virtualization, whether it's VMware vSphere-based or Red Hat Enterprise virtualization based in this case. You have a set of hypervisors. You have some management networks behind there with shared storage, in this case some block storage. And you've got a management system. So this is Red Hat Enterprise virtualization manager, which is based on the Overt project. But you could similarly think of that as vCenter. So this is your traditional data center virtualization. So inside that manager, you've got a history database with metrics for how your virtual machines are performing. So this is your traditional data center. This is great. You get a bunch of capex improvement, less cooling, less power, more systems running on a single piece of physical hardware. So then what would happen is, as the business grows, they realize that they don't only just want capex improvements, they want opex improvements. So this is your cloud based on data center virtualization. They deploy an appliance, CloudForms Appliance in this case, which is highlighted in blue there as a workload on there. And then basically the CloudForms appliance, which has these components inside of it, including something called the VMDB, basically would connect to the traditional data center virtualization to begin harvesting all that information. So it's pulling capacity utilization data, it's collecting events, and it's getting state relationship and configuration data from that deployment. So this is the state that we find a lot of our customers actually in. And now what they're beginning to investigate is, how do I start to leverage OpenStack? And so what they don't want to do is, in conversation with you, what they don't want to do is have a different management framework for reporting and collecting all that data. So what CloudForms allows them to do is, one, they could actually begin to deploy their OpenStack using Red Hat Enterprise Linux OpenStack platform. So this is for their cloud based workloads. So on the right, they still have their traditional data center virtualization running. On the left hand side, they begin to build out their OpenStack Cloud. So on the left, you've got your compute nodes and your controller nodes. And this is inside of their controller nodes, they might be running some of these services. So whether it's Solometer or the AMQP message bus or the Nova API, those are the particular areas that we hook into with CloudForms. So what they could then do is deploy another appliance on their OpenStack instance. So we deploy the OVF installation of this CloudForms appliance in both OVF. And we also provide it as a QCal file that you can load into Glance. And then they just launch another image on that side, and they can then federate their CloudForms deployments. They still have one UI they go to, and they can get reporting metrics across both their data center virtualization and their private Cloud based on OpenStack. So at the end of the day, this is kind of what it lays out to. So you've got Red Hat Enterprise Virtualization providing your virtualization layer, Red Hat Enterprise Linux OpenStack platform. And then you have Red Hat CloudForms providing the hybrid cloud management between the two. One thing I did want to point out, for those of you that were in some of the other sessions around OpenShift, the platform as a service offering, those run as virtual machines on top of infrastructure. So in this case, if you were running your platform as a service, you could see that the brokers and nodes, if you're familiar with the terminology, have open interfaces as well. And this is going to become important because as we move to kind of unify the operations management, we're not just going to go horizontal across providers from private infrastructure as a service clouds and virtualization and public clouds. We're actually going to go up into the application layer as well and begin to unify that. So what I'd like to do, we kept the, hopefully, kept most of the marketing to a minimum. Now we'll dive into demos. That sound good? All right, so the first thing I want to show you guys is adding a provider to OpenStack. So let me just flip to here. All right, let's take you through this. Adding a provider to CloudForms, I mean, right? Yeah, sorry, adding an OpenStack provider to CloudForms. So this is the CloudForms Management Engine UI. It's a web UI that runs on the appliance. When you log in, you'll notice you're greeted with a dashboard. There's a couple of menus up at the top. Under the infrastructure and providers, you're going to have your traditional data center virtualization providers, VMware vSphere and Red Hat Enterprise virtualization. But there's also a tab for cloud providers now that's been added in CloudForms 3. And so when you go to clouds and providers there, you're going to see a number of providers. There's an Amazon provider, so we've discovered a couple of regions. You'll also see there's an OpenStack provider that's already been discovered. So these are the two providers that we have. But if we wanted to add a new provider, we can simply go up to a configuration and select Add a New Cloud Provider. Under type, we'll select OpenStack. And basically, you could start to enter in some basic information. What user you want to connect to the OpenStack environment, so that's a Keystone user. And the host name that you want to use to connect to that, so this would be your controller nodes on your OpenStack deployment, an IP address. And that's basically it. On the right hand side, you'll see there are credentials. There's a default credential, and there's a AMQP credential that you can actually use to connect to the event bus as well. So you can validate the credentials. And once the validation is successful, you can add the provider, and you'll notice it's running there. So just with a couple of simple credentials, you can connect to the OpenStack environment. And then what happens is the discovery process will begin. And you could check off the provider and force it here. So you could refresh the information on the provider. And CloudForms will go out and spin up a new worker thread and go ahead and discover all the images and instances and all that sort of stuff. So after just a few seconds, you can see, well, there's still zero in there, because the job is running here. But after just a few seconds, the provider will refresh, and you'll start to see information in there. So it's pretty simple to add a new provider. So you'll see here, it discovered the availability zones, the flavors, the security groups, and the instances and images in there. All right. So the next one I want to take you through was reporting. So CloudForms connects to Solometer and actually grabs the metrics on an interval for what's going on there, and actually provides in-depth reporting of the instances running on there. And also, so this is pretty nice, because you can actually begin to report across virtual machines running on your data center virtualization, and you could report against instances running on OpenStack as well. So whether that's Rev or vSphere or even instances running on Amazon, you can begin to normalize some reporting across all those different environments. So you'll notice, here's my OpenStack provider. If I go and select the instances, there's two instances running, small OpenStack deployment here. And I'll select one of the instances here. So this is just an instance that's running Red Hat Enterprise Linux, but it could be running Windows or another distro of Linux as well. And then from here, you get a bunch of information about the instance. Oleg will talk a little bit later about the roadmap, and some of the things we're going to be providing in here from a drift history standpoint, and all this other stuff that's going to be happening that we already provide on the data center. But at the top, you'll see there's a monitoring button, and you can actually select utilization. And what's going to be presented here is a graphical representation of all the data in salometer that we pulled from the OpenStack deployment. So you'll see, in this case, I just added the provider, so I have to switch to hourly from an interval standpoint. And then I'll begin to see the average CPU usage disk and network as well. So this is just driving into a single instance to get all that data. And since we're gathering the events, you can actually start to correlate events to spikes on that timeline. So if you would right click on the disk IO when you see that little spike, you can actually say, show me all the events that happened that we got from the AMQP event bus at this time, and you'd be going to do some root cause analysis as well. All right, and then under reports, so that's kind of an active view of an instance. But if you would go to reports here, this is where you can actually start to drill into things such as instance and VM performance. So here's just a simple report that we generated. There's hundreds of reports that come out of the box, but this is one that I created called instance and VM performance. And basically, it's providing you with CPU utilization of both virtual machines running on Red Hat Enterprise virtualization and VMware alongside. You'll notice that there's instances as well in here that are showing. So my REL6-4 instance is showing obviously very low utilization based on the salometer. There's nothing running on them, but you can see the VMs right alongside the instances in that report. And you can also begin to generate OpenStack migration reports. So if you want to see the percentage of CPU that's been overallocated on virtual machines from VMware, you can begin to do that here. Maybe you just want to see ones that have been tagged as QA or development resources that meet a certain criteria. You can begin to develop those reports here. And just to show you, you can actually go in and customize all these reports. So there's ones that come out of the box, which are in yellow, and then the blue ones are kind of ones that have been custom developed. And just to show you the number of fields you can basically put into these reports, it's pretty amazing. All right. So that's the reporting side. The third piece we want to show you was ChargeBack. So while salometer provides you with a lot of data, it doesn't provide you with the ability to create kind of a rate table for ChargeBack and then apply those. And CloudForms has been doing this in virtual environments already and with public cloud providers already. But now we're adding this for OpenStack. So when you go up, you can select ChargeBack. And in here, you can actually begin to define rate tables. So you'll notice underneath here there's a rate table for compute and storage. If I select the default rate table that was set up, it's giving me an allocated or a usage rate table so I can charge by the CPU allocation or by the CPU used. And I can also do a fixed CPU cost as well and do some things with memory and network you can see there as well. And then I can assign them. When you assign them, you can actually tag. CloudForms has a very rich tagging implementation. So you can actually have everything that gets discovered in your OpenStack environment get tagged automatically based on various rules that you can set up. So in this case, I just did a default and just put everything under the default table. But if I wanted to, I could actually tag development workloads and QE workloads or different lines of business for ChargeBack. And you'll notice this is a ChargeBack report here. So there's two instances running on my OpenStack environment. I was able to take the salometer data and then apply a rate table to it to provide ChargeBack for my OpenStack environment. I don't think there was any other else on here. All right, the last one is self-service. So there's really kind of two concepts of self-service inside of CloudForms. And while Horizon provides self-service for OpenStack, some of the areas that it's lacking are really the ability to do workflow, quota enforcement, approvals, and those sorts of things. So that's really what CloudForms is designed to do. So if you select the Services and Workloads tab in the Appliance, you'll notice there's a bunch of VMs and instances running. You could filter them using the global filters on the left. But on the right-hand side, you can also go to Lifecycle Improvision. And just to take a quick stopping point here, basically this entire user interface is driven by role-based access control and tagging. So you can actually limit what the user sees. So if this was just a self-service user, I could limit him to not be able to see all these different virtual machines. He just gets a simple catalog interface, which we'll see later. But in this case, when logged in as the administrative user here, I'm going to be able to just go see all the different images in glance, templates in VMware, templates in Rev, and AMIs in Amazon that I have registered to this. But when I scroll all the way down to the bottom, those are all AMIs in Amazon. But you'll notice I have a simple image, REL64, that's associated with glance in my open stack of deployment in Boston. I'll select that and hit Continue. And then from there, I get a default dialog. This dialog is the default dialog that comes out of the box. But under that Automate tab, and we'll show it here, but you can go in and customize that entire dialog. If you want to limit it to certain fields, you can do all that sort of stuff as well. So in this case, it's asking for some things like an email address, first name, last name. Under Properties, you'll notice that we actually discover the instance flavors. So those will pre-populate automatically in here, along with guest access key pairs, floating IPs. So if you've got those defined on your open stack environment, I'll pick those up. It also does cloud in it as well. So if you want to have some user data injection via cloud in it, you can do that as well. So you can go ahead and provision that way. The other way that a lot of users are starting to move to is this service catalog approach. Here you'll see I just defined a very simple service catalog, and it's got one dev instance in it. And here you can actually go and order that instance just by hitting an order button. And then it's basically going to launch the same instance on the other end. But in this case, it's kind of a customized dialogue. So when the user selects order, it's going to just give them a asking for a first name or a last name. So customized dialogues with self-service quotas, workflow, all that sort of stuff inside there. Let me flip back to that. So that's a really simple demonstration. If you guys are interested too, feel free to stop by the booth at any time. And we've got these appliances running live, so we can actually show you a lot more in depth. I guess two last things to highlight. One is a lot of customers have existing CMDBs and event consoles, so this exposes a full web services API that you can integrate with as well. So if you have a favorite service catalog or existing IT operations management tools, the idea is this can be layered in between them to provide all that automation with the underlying platforms. So that's what it's meant to do. So from a strategic direction standpoint, before I hand it back over to Oleg on some of the roadmap things, this is really where we're going. We're basically providing these capabilities on the left. And we're really going to drive them across all the way from storage to infrastructure as a service to virtualization through the OS up to the development tools. We're fortunate at Red Hat to be involved in open source communities in all of these. And as Oleg's going to discuss, we're excited to be doing the same thing with CloudForms. So going forward, I just wanted to share what we're going to be focusing on in future releases of CloudForms. So last night, or this morning, we announced CloudForms version three. We're now planning the next release of CloudForms. And some of the items here will make it, and some won't. We'll have better clarity in the coming weeks and months. So the first thing is open sourcing. We're looking at open sourcing CloudForms, obviously. If any of you are interested in participating in that earlier or getting involved, please send us an email at that email address up there. And once we go live, we'll certainly let you all know. One of the other things that we're working on is integration with identity management, or the free IPA project that's out there, to allow us to integrate with various LDAP systems in a nice, clean way. We're also looking at adding more providers. So there's both cloud providers and infrastructure providers that we're looking at adding. There's Microsoft ones that we want to add, SCVMM and Azure. And we also want to go deeper on some of the other ones. We want to add more support in Amazon and OpenStack, which I'll be discussing in vSphere and Rev. We want to go both deeper and wider on the provider side. On the provisioning side, we're looking at integrating with Foreman. If you guys have heard of that project, that's one of the things that we're looking into. On the config management side, we're looking at integrating with Puppet and Chef. So you can actually do configuration management of the newly provisioned VMs through Puppet or Chef Scripts, but in a native way. We're also looking at integrating with OpenShift, allowing you full visibility into that and allowing CloudForms to visualize the gears and let you move workloads between VMs and gears. We're also looking at composite services. So you saw some of the services that James just showed about ordering an OpenStack VM. You can actually use that same service catalog to make composite services through your web apps or things like that. That's already in the product, but we want to start leveraging it to have specific types with a back end of heat and CloudFormations. On the OpenStack side, these are the things that we're looking at. So on vSphere and Rev, we actually have a technique where we look at the bits of an image. And without starting it or anything, we introspectively go through it and can read anything on that system. So for OpenStack, we're looking at doing image analysis. So if CloudForm sees an image, it can actually see what's installed on it, what package is. And once you have that visibility, you can see with the genealogy capability of kind of seeing what VMs are derived from an image. Now you can start seeing the insides of VMs or instances. Furthermore, for instances that live on Cinder, we want to be able to do the same kind of analysis of the insides of those instances as well. Once we have that information, we can start doing policies on it. We can start doing drift analysis, how things are changing. So it kind of allows us to do a lot of, opens up a lot of new functionality. In the current release of CloudForms 3, we didn't really focus on multi-tenancy. So that's one of the things we're taking back and we'll be looking at. A lot of customers have asked about software defined networking specifically when provisioning to make sure that we can provision an instance and set up more than its loading IP. We've done similar stuff already on Amazon where we started integrating with their VPC layer. We kind of want to do the same thing on OpenStack. In terms of Glance and Cinder, we actually want to give you full visibility from CloudForms onto Glance and Cinder to show you what's there. Once Solometer is able to emit events, not emit events, but provide APIs to give us events, we want to start consuming them rather than the way we're doing it now, which is by sitting on the message bus, which today is only in CloudForms 3. We sit on the AMQP bus. We're also looking at sitting on the RabbitMQ bus. But in reality, I don't think we should be sitting on any bus. I want there to be some API that we can consume instead. Finally, we're looking into, if you guys have seen the triple-O test car stuff, and if not, you should definitely stop at the Red Hat booth and take a look. We want to have CloudForms provide visibility into the resources that comprise the OpenStack Cloud so that in a private Cloud situation, you now have full visibility into your whole infrastructure that's supporting your Cloud. And through that, we can offer things like capacity management in terms of when you want to add new hosts, or when you want to add new storage, you can get a lot of interesting information that way as well. That's it. Yeah, so I think the takeaways are we've got the same capabilities that ManageIQ is bringing across VMware vSphere and Red Hat Enterprise virtualization. Since the acquisition, we are bringing those across as CloudForms. We're committed to open sourcing the code. We would love to, if you guys are interested in bringing operations management across OpenStack or any other platform, we'd love to work with you on it. So with that, we'll open it up to any questions. What was that? The tagging. Yeah, so the question was, how does the tagging relate to the charge back? Yeah, so the tagging relates to charge back and that you can assign the rates to things that are tagged with something. Yeah, so it's completely dynamic. So for example, on discovery of an environment, you can lay down tags based on what, in the VMware world, for example, what packages are on a specific operating system, you can then tag the system on discovery. And then those tags are then applied to the, are assigned to different rate tables. So it becomes really dynamic. So if a VM changes from development to QE to production in its lifecycle via tag, then the rate tables will automatically, the charge back reports will automatically update. You don't have to go through and statically assign those. Right. Any other questions? All right. Oh, all right. Oh yeah, yeah, the slides are actually, the best way to get them is go to my blog, which is my retirement plan, too. So go there, click on as many ads as you can and download the slides. I'm just kidding. All right, thanks. Cool. There you go. Only one question, only one half.