Listen to Nutanix Systems Engineer, Daeric Graeber, Nutanix's end-to-end consumer-grade management and desktop as a service (DaaS) solutions.
Speaker 1: On behalf of Nutanix and Carahsoft, we would like to welcome you to today’s podcast focused around Simplifying Operations: Datacenter to Desktop, where Daeric Graeber, Systems Engineer at Nutanix will discuss Nutanix's end-to-end consumer-grade management solution, Prism, and desktop as a service (DaaS) solution, Xi Frame.
Daeric Graeber: Okay, so yeah, so today we kind of wanted to talk about simplifying operations, right? Everything from the data center to the desktop, and how Nutanix can help in that journey. And we'll go over AOS a little bit about HV prism and then ran off the hour with the discussion around the Nutanix frame. If we kind of take a look here, right, we see more of what we consider traditional infrastructure, right? What we call three tier infrastructure, you can see at the bottom right you've got your storage environment. You know, this may be made up of a single sand, you know, maybe a couple of sands for redundancy. Maybe there's a Dr sand, right all connected into some sorts of switching environment, whether that be fine, fiber channel, you know, 10 gig, 25 gig ICO as a connections into the compute layer, right?
That may be three or four different Dell or HP or Supermicro servers, et cetera. On top of that, we have our virtualization layer, which is our hypervisor. And then lastly, ending with our application layer, right? That was where you're actually installing the O S and song applications that we're delivering to our end users and to our customers. So from this perspective, this can become complex. You can have multiple vendors in here, right? That storage may be made up of HPE storage. The compute may be from Dell, the switching environment may come from Cisco, you know, and then you've got on top of all that, you've got a separate vendor for your virtualization layer, and then obviously for your applications and, you know, OSA, right. So you can see, you know, multiple vendors in here.
So, yeah, so you can see here, right. You know, that traditional infrastructure, that three structure can be somewhat complex, you know, with the multiple layers involved there. So if we look at what Nutanix does, right, we can help you simplify your infrastructure with what we call ALS, right? So ALS is the Acropolis operating system. This is the, you know, the software that runs on top of your, the hardware that creates a Nutanix cluster, right? So if you used to take a look here on this diagram, you can see at the bottom, we still have our storage network, our storage environment, the two controllers there that controller one, controller two connected to a network up to our servers, similar to the last picture that we looked at, right. What Nutanix does with AOS is we essentially remove that storage environment and virtualize those controllers, right? So it's what we have, what we call a controller virtual machine.
Essentially what that's doing is in each of these nodes, within the cluster, that CBM is taking the, you know, whatever storage resources are available to that node and presenting it back to the hypervisor, right? So if, you know, if no one has a mixture of flash and spinning drives, right, it's going to take that storage, combine it with whatever's configured in node two, three, four, et cetera, and ultimately make that, that one big storage pore that distributed storage fabric across your environment. What this allows you to do then is be able to scale, right? So every time you need to add a new node, whether that be, you know, a storage, heavy node or compute heavy node, that can be scaled independently, as you add new nodes to the cluster, it will simply take that, create a new CBM, take the storage resources from that TBM and present it back to your storage.
For right now, you've got this distributed storage fabric across all of the nodes and time the environment compute and memory resources are the same. They're pulled based off of what's in the node. All of the notes don't have to match. You don't have to have the exact same processors, the same memory configuration, the same drive configuration, pretty much the only guard rail that you need for this in regards to, you know, requirement is that they all be on the same manufacturer, right? So I mentioned a couple of times this can be the physical hardware. It can be, you know, Dell OEM that can be an ex appliances, or it can be HPE. There's a couple multiple different vendors that we work with, but to create a cluster, you have to, they have to all be of the same OEM. That's kind of the only requirement, six months down the road, after you implement Nutanix, you know that you've got a storage requirement, that's going to require quite a bit of resources.
You could add a note to that cluster that was, you know, filled with spinning drives, and maybe just a little bit of SSD and, you know, the bare minimum processors and memory. And then that would automatically increase your overall storage resources across the entire cluster, right? Circling back to that CVN. And what that takes care of is that takes care of the data, tiering, the compression and deduplication, RFQ, RFP, and erasure coding. So with Nutanix, you know, we don't deal with rain or lungs or anything like that. So if you look at what we have RF two and RF three that's redundancy factor of two and redundancy factor of three, essentially every time a application or VM writes data to it's disk, it creates a secondary copy somewhere in a cluster. If you've got that configured with RF two, if it's RF three, that would make three copies across the cluster.
If we were always redundant, right? Erasure coding, we do support that. That is more of a, you know, where we're striping the data across the cluster that you typically use for specific workloads, right? Maybe it's a large video repository or something like that, large storage repository that you need to kind of maximize that storage within a cluster, that's where you would leverage the erasure coding. So in addition to this, you know, something that sets Nutanix and AOS apart from, you know, some of the, the market there is we have what we call data locality. So if you, if you on note number one, and I'm writing data to disk, that VM is going to write to disc on the number one, right? So it's going to keep that data local all the time. Again, based off of RF two, and our free will keep other copies inside the cluster as well.
AOS is by default going to keep that IO local, that way you're always getting the best performance, right? So if you have a failure on node, number one, that VM is going to spin back up on wherever we have the secondary copy and the cluster took to maintain that data locality. So we'll talk about deploying Nutanix, right? You know, a lot of times there's a lot of planning a lot of anytime you deploy a new environment, new infrastructure, you know, it's usually weeks of planning, installation, all of that with Nutanix, we have something we call foundation, you know, once the nodes are racked and cabled work together, it's as simple as logging into a web interface providing it a few pieces of information, right? DNS, IP addresses, et cetera. And then you start what we call our foundation process, right? Typically within three or four node cluster, you can do that in less than an hour.
You know, you can do multiple clusters at a time. You can do clusters at a remote location. And if you say, if you have like robo offices, et cetera, it's the same deployment process for all hardware for all hypervisors, right? No matter if it's Dell hardware, HPE hardware, Nutanix hardware, and you have that. And if it's, whether it's the HV hypervisor cluster, hyper V or VMware, all of that's taken care of by our foundation process and takes it all the way up to the point where you're ready. You know, once that foundation is done, you're ready to start creating VMs and move forward. So with our prism interface, right, prism is the the kind of the the web interface based in Monaco web interface that you access your Nutanix cluster with. Now, you can see here prism central is sort of what we consider it to be our manager of managers.
Right? If you have multiple clusters in your environment, you have one place to go to create the NS in any of those clusters to do maintenance, to do upgrades. Any of that stuff is all taken care of my one interface, right? So if you're used to having one interface for your storage, one interface for your hypervisor, one for your compute, et cetera, all of that is managed under one, single as interface with prism central, the planning and automation with prism central, you get a lot analytics about your current environment, right? So once this, once your Nutanix environment runs for, I think it's, depending on parameters could be up to about three weeks, it will start giving you data and insights about your environment, right? So it can tell you that if you stay on this, the same path that you're on in regards to growth and additional BMs or stores that you're adding, it'll actually give you a runway of when you would need to add additional resources.
Right? So, and what's nice here is you can actually go in and plan for things like that. So if you know, two, three weeks down the road that you're going to be deploying, you know, 20 or 30 new VDI users, or maybe a new application, that's going to come along, it's going to require three or four different VMs. You can actually go in there and plan for that. And it would show what the impact would be on your current attendance cluster. And if you would need additional resources to be able to handle that, right. Additionally, you know, as, as I said, it runs over over time. So it will actually report back to you and let you know, if you have VMs that are over-provisioned under provisioned, maybe inactive just turned off there, but you know, still reserving the resources, right? So if you have a VM that has two VCPU and four gigs of Ram, and it's constantly maxing out at the, you know, it will, it can actually send you alerts or send you emails to say, Hey, this VM is, you know, constantly running it full tilt and a hundred percent of its resources.
And you can actually make a decision there to increase on the fly. You can actually automate that as well. Right? So if it's something that you don't want to worry about, you know, you know, I've got this VM, I'm going to only assign it to a VCPU at four gigs of Ram, but there may be points in time and where it's going to need more than that. I never wanted to use over say eight gigs though. You can automate that and allow it to increase, right? So if that contention does come into play, it would send an alert, automatically increase the memory for that VM and then, you know, you'd move forward, right? No intervention from your end, but we'll talk a little bit about AHV I mentioned earlier, you know, we give you a choice of hypervisor. So AHV is our Nutanix it's stands for Acropolis hypervisor.
Probably the best comparison that you could think of here would be if you're familiar with KVM within the Linux environment, you know, a lot of the major hyperscalers, Amazon GCP, they use KVM as their underlying hypervisor. We kind of took a page out of their book and did the same thing, but if you've ever deployed KVM, or if you're familiar with it, or you'll know that it's not, you know, there's multiple packages, you have to install, there's different tools you have to use to manage it. It's not the user experience that we would expect everybody to want or want to have, right. It can be quite cumbersome. So what we've kind of done has, you know, have added that feature set that makes it more adoptable, easier to use for end users. You can see here, it's got all the same enterprise ready features.
One thing to note here, though, it is performance optimized for Nutanix and for HCI, right? So if you think the hypervisors like VMware and to, you know, hyper V these were hypervisors that were created and built before HCI was really a part of the market, right? So what's KBM. And what HP specifically is it's optimized to run a, in a hyper converged infrastructure, right? Everything else, they're, you know, things you're probably used to, right. And CPU automated CPU about compatibility affinity rules, all of those things included, right. For us at Nutanix, we don't worry as much with what the hypervisor is, right? We're, we're ultimately here to deliver an experience to end user that we use the analogy of when you go to AWS or to GCP, to spin up a VM, you don't select a hypervisor that you want to choose for that. Right.
You know, they're going to just give you a VM with whatever left that you want on it, and you'll deploy your applications. You don't at any point request that they, you know, just be running on hyper V you're on VMware. Right. And that's kind of the way that we look at things too, right. Hypervisor just seems to be more of a commodity within the data center. These days. I mentioned a little bit about this earlier, but you know, with prism central, you can upgrade patch your environment, simply with one click and upgrades that we offer. You can use that to upgrade AOS, prism, files, objects. If you're using more of the Nutanix full stack, HPE as well, that you actually can upgrade hyper V and VMware as well, also through, you know, there's no need to log into V-Center upgrades, your hypervisor. You can do that directly from our prism console, as well as upgrade kind of the low level firmware, right?
So your bios, your SSD firmware, or anything like that, the controller from where any of that stuff is all managed within for prism central through our LCM, right? Our lifecycle managers, regardless of whether you're using Dell, HPE, Nutanix, and whatever hardware, you don't actually need to log into say an hydrant card anymore, or an ILO, you know, or some kind of other IPMI to do low level firmware upgrades that can all be handled through prism central and the life cycle manager. So Nutanix frame, right. So we'll start a little bit and just kind of round out with Nutanix and user computing solutions. So on the left, you know, obviously with Nutanix, you know, when clinics first started as a company, we are a very Bible. So option for VDI in general, right. I mentioned earlier about data locality and how that can help. So you can imagine how that would it help you with performance in a VDI environment.
So, you know, a lot of our first deployments in first customers were VDI users with either Citrix or VMware horizon, you know, and we look at those more of like a traditional on-prem VDI solution, right? Which we still support. We still have, you know, many happy customers using our solution for their traditional VDI environment on the right. We have what we call Nutanix frame, which is more of like a hybrid desktop as a service solution. And we'll jump into that now. So Nutanix frame, you know, essentially the cloud platform that delivers, you know, it says windows app, this can also deliver Linux applications and desktops to users over just an ATM all five browser. Right? So if you look at the top here, you've got what we call the frame, the frame control plan of frame cloud by fact plane. Essentially, what you can do with frame is it's just the broker.
If you are familiar with VDI and you're familiar with using Citrix or VMware, you know, frame is our broker, right? That's what we host out in the cloud for you you're ready to, to log in, to spin up, you know, new VDI instances, as far as the infrastructure grows, you can choose whether it be AWS, Azure, GCP, or on-prem Nutanix, right? So we kind of round the circle out here, right? The way that the, you know, the process kind of flows is you pick your infrastructure. Where would you like the VDI instances to live? Where would you like the users to, you know, to be connecting to whether that be one of the three hyperscalers or AWS, sees how you're going to authorize those users, but maybe it's active directory or some kind of a SAML authentication, Okta, something like that, choose the applications that you want to deliver to them.
It can be Photoshop. We, you know, we do very well with high-end graphic, intensive applications, like AutoCAD, Photoshop, those kinds of things, Adobe software, how you want those users to handle storage, right? Typically with a frame environment, we don't see too many persistent desktops. So to say, you know, if I create a new photo file, how am I going to save that? We have tight integration with a one drive, Dropbox, Google drive those different things, as well as like on-prem SMV shares. And then Nutanix files is our offering for that, and then how you're going to deliver it. Right. So I mentioned earlier, you know, deliver over an HTML five browser, as long as they can access that they should be good to go. So you can see here kind of a little bit how that looks to an end-user, right. If I log in that's, they're going to be greeted with, this is actually an application launch pan, right?
So I'm logging in, I get access to these different applications. I can click on one of those. And it's going to boot that up into sort of a sandbox application where you don't have access to a full desktop. We also do have the option to deliver the full desktop experience. It just looks a little different when they click on it, it would join into the desktop. They would, then you would say that like, you know, a windows 10 windows, 2019, whatever OS underlying iOS, that's the desktop they would be greeted with. Right? You can see different user features there, obviously all of the main ones, 4k monitor support frame does really well with low bandwidth requirements. Based off of our proprietary protocol, you don't have to have a gigabit fiber connection to be able to utilize frame, right? You can utilize the frame, desktop, the frame application environment, you know, over something like an LTE connection, right.
Helps out definitely with some of our K through 12 students right now, because everyone's working from home remotely, maybe have limited bandwidth are still able to log in. You know, even if they have just a Chromebook at home, they're still able to log in and actually launch and drive, you know, some applications that may maybe even be GPU intensive. Right. A little bit about security, right? Obviously VDI in general, you know, one of its biggest feature sets, there would be the security, the fact that the data is not living on individual desktops, we have recently been FedRAMP authorized with Nutanix frame. You can run frame instances in Azure government, AWS, Gulf clouds, right. We, you can see all of the different compliance so that we can all the different certifications that we can comply with there in regards to the frame and delivering the end user desktops.
So your users, a couple of different things here, right. So I mentioned earlier that you can choose your infrastructure, right? Obviously a big use cases you have, we see user in a lot of the time, especially during the, you know, in the past year or so, where they needed to be able to deploy a large number of users almost immediately, right. To be able to give them access to resources and different applications. So we have, you know, that burst and capability where we can, you know, maybe connect a AWS instance to your frame broker that would help out well, while, you know, in the meantime, maybe we would be looking at more of a long-term on-prim on a cluster for, you know, to help mitigate some of that cost in the long-term. So you can see, obviously there's different advantages. Public cloud obviously has the, you know, the cost that comes with it, right?
So if you've subscribed to frame you, you know, you would scratch subscribe to a frame licensing from Nutanix, but then connecting that to a hypervisor that would be, you know, they would also be an additional cost there. I'm sorry, connecting that to a hyperscaler, there would be an additional cost there. Right? The nice thing there is, is that again, it's a, you know, very personable, very easy to deploy and very quick to get those desktops up and running private cloud that's, you know, again would be an on-prem Nutanix cluster that you would connect your frame broker to, and as users log in, they would get connected back to a VDI session on prem in your on-prem data center helps out, you know, obviously in certain use cases, cloud has some limitations in regards to, you know, GPS that they have available. Typically, if you have, I'd say moderate graphic applications, that would be fine for a cloud environment. But if you are really trying to deliver a specific kind of use case or specific specific variance to an end-user that may require an on-prem cluster, right. Where we can actually add additional GPU resources to that on-prem cluster.
Speaker 1: Thanks for listening. If you would like more information on how Carahsoft or Nutanix networks can assist your organization, please visit www.carahsoft.com or email us at nutanix@carahsoft.com. Thanks again for listening and have a great day.