 All right. Hello, everyone. Good afternoon and welcome to Listening to us about OpenStack and VMware. And my name is Hari Kanan. I work for VMware and I'm joined here by my colleague, Jao. Jao, do you want to introduce yourself for a second? Hi, my name is Jao and I'm a tech marketing engineer at VMware. Okay. So we'll spend the next 30 minutes with some slides as well as some demo to show what VMware and OpenStack really means in practice. Before we get started, maybe with the show of hands, how many of you actually know that VMware has an OpenStack distribution? You know that. Okay. That's good to know because some of the presentations I've done in the past in a few other places, people didn't know it was one of a fairly well kept secret. And the purpose of this show is primarily to bring out the value proposition that OpenStack on VMware stack is actually a viable and productive option that customers already use today. So that's kind of the objective for today's presentation. So some stuff that we'll be discussing are all forward-looking, so standard stuff. We would like to make sure that, you know, some of the things that you're hearing may not ever show up in a product, but if you have any questions, feel free to reach out to us offline. So what is VMware's OpenStack? We call this BIO, VMware Integrated OpenStack. So the first thing is that I want to clarify that it is actually a standard OpenStack distribution. There is not a specialized version that is somehow deviant from traditional OpenStacks in any fashion. We do support its, it used to be Defcore compliant and actually now the Defcore name has changed to interrupt. So we are compliant with the, including the latest standard 2017, interrupt 2017 standard. So it is actually a compliant version of OpenStack distribution. And what we do is we take pretty much the upstream OpenStack and then we add our own management capabilities, both in terms of the installation, reducing the installation complexity, day zero challenges, as well as the day two challenges in terms of both operationalization as well as upgrade and migration issues. So that's kind of our approach to OpenStack, but from an upstream perspective, we are as compliant as any other OpenStack distribution is. And the goal for us is to run this on VMware's stack, which is basically vSphere and NSX and vSAN, but vSAN is not a key requirement. NSX is, most of our customers use NSX as the SDN provider, but technically we don't require NSX also, but it's the most preferred suggestion. The minimum requirement is to have a vSphere environment and that's the only platform we support. We don't support other hypervisors like Zen or KVM. Again, following the theme of what we support and what we don't, we have the standard projects that we support, Keystone, Nova, Cinder, Glance, those are some of the products that we actually support out of the box. And we take a distributed control plane, we run on vSphere, both the control planes run in HA mode and we support compute nodes which are actually vSphere clusters and resource pools are exposed as compute nodes and we have written the VC driver. In fact, it was a community-driven, one of the early contributions, a significant contribution for the driver actually came from HP ironically and then we have, as VMware, we have also enhanced. So it is a community-driven project, pretty much everything we do, we push upstream. So we don't really keep a lot of fold-off versions that anything that's very proprietary to us, work with the community and push that upstream. There may be obviously some timing changes because we may have productive customers who require faster patch, so we do provide paths and sometimes reviews take a little time but we work with the community to eventually check it into the upstream release as well. So our strong belief is OpenStack provides a great abstraction. It provides an IS platform but while you might have heard this cliched word, pets versus cattle and OpenStack kind of platforms are built at least grounds up to support the cattle type of workload where applications are dealing with a lot of the resiliency and there can be failure and the infrastructure and applications know how to recover from those infrastructures but at the end of the day as an administrator, as someone who is managing an OpenStack environment you still want to make sure it is running on a robust platform and that's what is the single biggest value proposition that we bring to table. We are providing OpenStack semantics, we are providing an OpenStack distribution exactly the same distribution that you get from any other vendor on any other hypervisor but we are providing it on a battle-harden tested infrastructure from a hypervisor and a networking perspective, vSphere as you know is still the dominant hypervisor platform in the private cloud context and our goal is to make sure that our customers have a choice and one of the choices we offer is an OpenStack platform. So our goal is to support and we believe that you know you have, if you build your IaaS platform on top of a stable battle-harden tested infrastructure platform then the result is going to be a very stable version of IaaS platform and then you built on the vSphere platform then you are able to leverage a lot of the benefits of the platform such as the VMware's HA capability, DRS, vMotion all those things that have been working in a productive environment for ten decade are readily usable in the OpenStack context so you get sort of the best of both worlds in terms of both having your OpenStack from an open source community perspective as well as having a battle-tested infrastructure perspective. Same with NSX and some of you might have been in the presentation earlier from some of my colleagues but NSX is actually a dominant SDN player. We got this through an acquisition that we made many years back the Nisira guys and we actually have, Dimitri is laughing here and we have been part of the open source and OpenStack community from very beginning having been one of the founding members of the Neutron, what used to be called Quantum Networking many years back so we have a legacy, we have a huge IP that we have contributed back to the community as well and that's the NSX is a key capability for us and you bring all the value propositions of NSX like micro segmentation and also the other capabilities like load balancer and firewall all of that can be brought into the OpenStack fold built on a stable platform and then the storage is again, since we exposed Cinder as the VMFS volume for us every single device, vSphere has probably one of the largest hardware compatible devices we probably connect to every single device in the world, I mean storage device and because our Cinder driver is actually based on VMFS you automatically benefit from using any of those hardware device as well as any software defined storage capability as well very broad and also deep integration with most of the infrastructure components and this is a very important feature I used to be actually many years back in HP's OpenStack team and early in those days 2012-2013 we were one of the very first people to have stood up public cloud based on OpenStack, I think it used to be the SX Diablo or SX days it was a technically challenging feat, somehow the team managed to stand up a public cloud and a lot of our enterprise customers saw that it was actually a great environment for them to benefit similar to the AWS environment what they did was as OpenStack started becoming the de facto private cloud software the most often asked question was hey we like what you have done from standing up a public cloud based on OpenStack but I would like you to replicate that so give me a distribution of that public cloud that you're running for my private data center that's where actually a lot of the challenges and it's still a challenge for multiple organizations even after three years has been one part is the distribution but the operationalization of OpenStack like any other cloud environment is probably the most complicated thing that you can actually as an admin you can face which is still a problem but with our attempt to integrate this very closely with our industry leading products like V-realize operations and V-realize log inside we are able to bring that integrated view of not only deploying on a day-zero basis but also troubleshooting with various integrated solutions and some of it which Shah will demonstrate today we are able to bring the value of a monitoring system into a productive environment so that you can troubleshoot and keep the cloud running and which is pretty much a lot of our large customers are actually using today so the integration of monitoring both from a monitoring and logging troubleshooting is a very key aspect of the suite that we are selling the vios product that we are selling to customers because it's not just about deploying an OpenStack environment but how unified it is from an operationalization perspective is probably what is going to make or break the success of your private cloud initiative so to summarize the benefits of YOVM vSphere integrated OpenStack the key value proposition for us is it is delivered as a single OVA it's extremely easy for us to install you have a few choices to do the installation one you can do a compact mode installation if you want to do a POC or even a small scale cloud it's very easy for you to install you can install all the OpenStack services in a single VM or you can have a scaled out installation which the size would be around 7 VMs for you to do the deployment for three nodes for your database for quorum two nodes for the OpenStack services running in HA mode and two nodes for load balancers so you can choose whichever way you want to install and you are able to take advantage of the native vSphere capabilities like HA so that if you have a failure in control plane you are able to restart with minimum downtime so that's value proposition the installation for small installation if you have an existing vSphere environment that's already set up vSphere and NSX the installation we have had customers who POCs that we can get started in as little as 30 minutes if you have a scaled out installation it takes anywhere between an hour to hour and half it's as simple as actually it's that simple to get yourself up and running in a private cloud the number one value proposition we bring is we have sort of tamed the base we actually can get you up and running anywhere between 20 minutes to an hour and half depending on the scale that you install that is a key value proposition for us the second would be the ability to have an integrated view of your both operations data operations in terms of troubleshooting monitoring through the integrated view point views we get from Vrealist operation in loginset the third value proposition for us would be your updating and patching that's actually a very seamless process many other deployments that I've been doing this for more than 4-5 years stumbled in update process most customers we feel like they are not able to move from kilo or even older versions to a newer version of OpenStack because it's just too complicated for them to move up so what we have done is sort of industrialized that process and streamlined it where you can actually have a blue-green upgrade where there is really zero downtime for your workloads or your apps and then the installation process it takes anywhere from half an hour to an hour hour and half depending on the size of the cluster that you have one of the biggest advantages is we don't deploy any agents on any of the worker nodes or compute nodes and that means we have nearly a zero downtime for not nearly actually zero downtime for your applications because there is no agent to update and there is no services to update on the compute node that might force you to restart or there is a disruption of the services restart the node or the service itself so it is pretty much zero downtime during the switchover from a blue to green there will be some interruption of the control plane itself which means you may not be able to do some operations on the control plane but the workload should still be running without any downtime so that's actually a very important advantage sometimes we actually have customers who do the upgrade from 2.5 to 3.0 and we don't even get to know that so it's as simple as it's a self upgrade process you don't require any professional services that's the other fourth value proposition I would say which is most of the customers do not actually even engage professional services to set up a cloud like a week or two professional services engagement it's a simple installation process if you are a vSphere administrator you know how your vSphere environment works most customers just do installation by themselves there is no need to engage any complex services engagement for that so this is another important question that I keep getting okay what is VMware why is VMware interested in this because I came from an environment where it was either VMware or OpenStack they were not treated like it was never a question of VMware and OpenStack for many customers it was that question always used to come in keep coming back to this question of why is VMware interested what is VMware is it committed in a longer term are they just doing this actually there are a number of press releases some of the largest customers who have issued supporting statements for us we really believe OpenStack is a strategic opportunity for us from VMware perspective and we actually while there is a general questioning and many customers many companies might have HP exited the OpenStack business people are thinking oh okay does it mean OpenStack is either maturing or slowing down every company had made a decision to either continue on or exit OpenStack for their business reasons in our case we actually have double down on OpenStack this year we have made significant investments to go after the NFV market so we have put R&D and engineering resources in supporting OpenStack for the NFV use cases and that side effect of that is actually our core OpenStack platform is going to be a lot more hardened and battle tested for the NFV not only the NFV use cases but also for traditional data center use cases so NFV because NFV is a new investment a new opportunity for us we have made significant investments this year in terms of engineering resources second is strategically NSX is a very huge growth business for us you might have heard from my colleagues Eve and Dimitri on how it is changing perceptions and how it is actually accelerating innovation in our customers so NSX is a key enabler for us for VMware it is a growth business and OpenStack is a significant enabler for the NSX business pretty much all our customers use our version of OpenStack with NSX and that's very important for both NSX business as well as the OpenStack business within VMware then we have our own CMP or cloud management platform which is more of a governed environment to run a different type of applications more of IT managed catalog, portal to deploy a VM and other services we have that called viriles automation and viriles suite but increasingly we are seeing a need to integrate both the OpenIS platform with a proprietary viriles cloud management platform for us to marry the best of the both worlds one is from a governance perspective people like the high governance IT managed environment but still want to expose a open API standards for our developers so we see an increased interest in blending those two and hence we see there are opportunities some of you might have been in a presentation that we made a few hours back we are also using OpenStack as an underpinning technology for building a new Kubernetes distribution as well so that's another growth area for us as the world moves towards containers and adopting Kubernetes as a standard we are leveraging our expertise and on OpenStack and building a Kubernetes distribution and few other things like I already alluded to this we realize competitiveness is very key for us and OpenStack is actually a strategic, plays a strategic role in that and then because we are now part of the Dell EMC family we are seeing lot more opportunities for us to build an integrated solution to go to our customers integrated hardware and software solution I'm going to quickly run through how much time I have we have 15 minutes left oh totally? okay so I'm going to quickly run through it this is the last thing probably I want to go through in terms of the slideware and again I want to make sure that the key takeaway for us is not only that we have a viable, vibrant distribution of OpenStack but we are also contributing heavily back to the community both in terms of OpenStack as well as other open source projects VMware, I would be the first one to admit is not well known in the industry for their open source contribution but that image and actually both the image as well as the perception as well as the contribution is changing dramatically we have not only started contributing a lot to open source I mean OpenStack but we are also contributing heavily to other open source projects in the container space so you'll see more and more open source contribution from VMware and that hopefully will change the perception that VMware is not just building proprietary software but we actually have a number of open source related contribution that we are making back to the community again some numbers as I was saying I am running out of time so the last thing right I mean it's not just a fly by night thing for us as you can see we have actually a track history of track record of having OpenStack initiatives within our company for a long time as when we bought Nesira we had already been major investors in OpenStack contributing to both OVS switch which was pioneered by the Nesira team as well as the Neutron used to be called Quantum at the time we have been contributing to it and then in 2014 we launched the beta program but early in 2015 we announced our first version of OpenStack and we have continuously improved on that OpenStack we have now significant number of productive customers that are running OpenStack so we have been releasing and refreshing our upgrade our releases to closely tie with the upstream the current version 3.x supports Mitaka based and sometime this year during summer hopefully we will be releasing and refreshing our product to a Okata based release as well that's our aspiration so we have been catching up with the upstream pretty consistently one major release to refresh the upstream release and then multiple one or more minor releases to add some features as well as bug fixes, customer requests and so on so last thing I want to point out here is from a roadmap perspective we have a very steady pipeline this summer as I mentioned we do plan to release our Okata based release and then that will bring in bring with it a number of data center related features like supporting multiple vCenters that would tremendously enhance the scalability of our product and also putting in a number of features to support the NFV use cases again as I alluded earlier the NFV use cases are sort of a win-win for us one it attracts the NFV we are able to now go with a credible product to the NFV market but also hardens our existing data center in terms, in ways that we haven't pushed the product in the past so last thing I will say and then leave time for a demo from couple of demos from Zawis it's not just me standing here and preaching that VMware is actually an open stack friendly company it's actually a media is one of our largest customers pushing us in really larger scales in multiple sites just issued a press release they have been a happy customer for nearly two years now they use open stack from Vio from us and also HeadServe is another I forgot to put the customer name but it's HeadServe it's another service provider big customer of ours there are number of other customers some are publicly referenceable, some are not if you have any questions or customer references I see some customers already here which I don't want to identify at the moment but we do have customers who have been using it at scale and we are growing at a fairly fast clip at this time Zaw, do you want to take over from for the demo? So I'm going to save some time for Q&A as well so I will try to go fast 15 minutes I think we won't have 10 minutes left So like Harry talked about Vio really has simplified the entire open stack experience you can deploy both a full scaled out version of open stack we were talking about in the magnitude of minutes versus days and weeks and I want to first show you how simple it is to deploy open stack using Vio and then I'm going to show you some examples of how do you actually manage it once it's deployed and lastly I want to show you guys how we can seamlessly upgrade from one major open stack release to another release without experiencing any sort of a data plane outage during the process So in the interest of time So the first demo I'm going to do is from Vio it starts with a single OVA All you have to do is go to vcenter load the OVA and everything you ever need to deploy that environment is already bundled So simply put in a few IP addresses and just a few clicks you're going to have a fully operational open stack environment without having to worry about package management without having to worry about configuring servers making sure the storage is online everything is bundled into the process already So let me first demo that video So basically this is the process the first thing you do is you log into vcenter and you go to the host and cluster view you say okay I want to deploy a new OVA file and you select the OVA you want to deploy and it's going to ask you some questions about permissions and then you assign a name to where you want to put this deployment and select the cluster where you want to put that particular VM select the data store that's corresponding to it and select the networking segments for that particular server and put the passwords this is what I was talking about enter a few IP addresses so that when we boot up the open stack controller VMs we know what address to assign and this is also going to be what we use to build the inventory file for our deployment automation so that we know which VMs to do what so this is part of the IP address for inventory file as well and once you do that you are ready to go the VM boots up and then you can proceed to deploy the open stack by just clicking the deployment button so you can select the configuration file basically is where is your NSX the IP address of the NSX what is the IP address of vCenter it's navigating you through the process just to confirm those settings are correct and then this is what Harry was talking about in terms of the number of VMs so I'm deploying a fully scaled out configuration with 7 VMs and that's it, its open stack is deployed now the moment that deployment run finishes you can log on to Horizon Dashboard you can create projects you can assign users to the project and we also bundled images just to simplify the experience so you don't have to go especially if folks are new to open stack you don't want them to figure out how to load an image in there and take hours just to do that we also have images already bundled in that's fully operational all you have to do is just create some networking and assign the hostname and everything it will just work so I'm not going to... I think most people already know about this so I'm just going to skip ahead and move to the next video so the second video I have here is the let me just give you some reference point Harry talked about the day to management so it's just because you have open stack doesn't mean that you have full visibility into your environment what kind of error messages am I having am I having capacity issues do I have how much is actually costing me to run this open stack environment so all those information that you need to understand the deeper visibility into your environment we actually pre bundled into the bio release itself all you have to do is load couple plugins from our market exchange and you have everything you need to do the day to support and let me just quickly do the videos I'm not going to go through everything just because of the time I do want to show you guys the upgrade process because that is something that we believe because of the unique architecture we have only the vCenter with the open stack approach let me just talk about this real quick so this is for you to do day to monitoring all you have to do is go to our marketplace exchange download the plugin for open stack, install it set up a couple credentials about how you log into NXSV how you're going to log into open stack using the admin account and then what type of SNMP attributes if you have anything that you need to manage there by doing that we automatically populate a set of inventory that you can look at so the first thing you can look at is open stack controllers every single, the seven VMs I deployed earlier the corresponding process that matches each VM we manage that as a single entity you can draw into each type of virtual machine and figure out if the process is running as expected and what process is running where so in this case we're looking at the nova compute so only the nova process is running and then you have the controllers which runs the memcache to enable faster authentication of tokens and then we have glance running on the controller and we also have nova as well as the sender from servers is running all the controllers so let me just fast forward a little bit and we also have the database node this is where the sql is running as well as the mq processes so finally you can also we also manage the the infrastructure components as well it's not just open stack we have because we are building this on top of battle proven technology we already know how to manage the resources underneath so you get visibility into your ESXi you get visibility into your NSX controllers so everything that from the API level all the way down to the physical plumbing we provide you with every single data metric that you could get prior to open stack so you can get it with open stack deployed probably let me just fast forward and see if there are any questions and then we can go probably to the upgrade let me go to the upgrade as well so are there any questions that we can answer in terms of open stack please use the mic that might be useful I know that at least for some smaller well enterprise I'm thinking of smaller carrier service carrier that I'm talking to one of the reasons they want to get off VMWare for the reasons you pointed out because right now their services their unified communications as a service is so SIP driven and they want to get off VMWare because they see this sort of a emerging revolution occurring with network function virtualization and BNF and so on and they know it's happening on open stack so they want to get it but they're terrified of open stack for the obvious reasons and definitely the upgrade a big deal to them but one of the reasons that they want to get off open stack well too one frankly is money because I believe your licensing costs have gone up recently in the last year I think so I was told and the other reason is open I mean open is really the new currency I think it seems to be a trend even bigger than just software it's almost like a new economy so just wondering in terms of that let's look at costs first do you see any will there be for somebody going over to VO say a medium sized environment would there be a cost saving or are we just looking at a sort of integration of open stack costing good question so definitely for us honestly so far in every customer engagement we haven't had cost to be a major issue people have a perception of VMware versus open stack that has been a challenge but in terms of cost it has not been because if you really look at cost cost has to be looked at in a lot more holistic fashion both from a licensed cost and also ongoing cost and so on so our open stack distribution is actually it's as a free entitlement to our existing customer base vSphere if you have a vSphere then you get it for free we don't charge the open stack distribution itself we just pay a very nominal fee for support some $200 list price or something so very nominal fee so the cost for the open stack distribution really comes from the two key components one is the NSX and one is vSphere but I based on all the research that we have done if you put the stack together compared to any competitor we are actually on par or sometimes even cheaper if you have an ELA with VMware and so on so there are ways to mitigate even the comparable cost that we have with every other distribution as such but that's only from a day zero cost perspective right but then you have to think about the rest of the cost which is how many people we have and what is the outage going to cost you in terms of if I have a downtime of one hour what is that going to cost your business that's where I think the value of vSphere which is really tested infrastructure that kicks in where you now can actually with a straight face we can justify that the investment we have made in vSphere and the investment we are making in NSX makes the platform a lot more robust so on an ongoing basis we are absolutely confident that you will save over the long run it's the same rhetoric narrative we keep hearing about public load versus private load yes if you are running a private workload on public load as a bursting or a temporary use case it absolutely is a lot more cheaper as opposed to keeping a server idle in your data center and the power and cooling and the people resources you want to put but a study after study has shown that if you run 365 days a year a particular set of workloads on a public load it's actually a lot more expensive than running it on a private load the same kind of argument extends into this scenario also where if you convert the cost over not just installation cost which we believe are very comparable then we actually are fairly competitive in terms of that we are absolutely a lot better off in terms of the recurring cost for you running your open stack environment so we are convinced about that okay so I've been told to cut it but we are going to be around happy to take questions on the sidelines the same demo I'm going to show you here if you guys stop by A2 I'd be happy to show you guys the things that we couldn't show you today because of the time the upgrade I really believe that it's the way we are uniquely because of the separation of control plane and data plane right so you could do whatever you wanted to do but move between major releases I'm going to show you guys if you stop by the booth how we actually maintain two separate control plane briefly and then flip over to after everything you test everything is working the way you wanted to then do a cut over right everything is controlled so stop by I'm happy to answer questions on the sidelines if there are any more thank you for your time thank you