47. Virtualization Module Introduction
I think you would have to agree that if there’s one technology that really propelled cloud more than anything, other than, of course, the Internet itself, right? We know the Internet is critical here, but what’s the thing that came along and really flourished post-internet creation that enabled all of this? It’s definitely virtualization. Now, remember, virtualization has been with us for many decades. I’m not suggesting that’s some brand new technology, I’m just saying that, boy, when virtualization really started to flourish over the last couple of decades, that gave great rise to cloud tech. In this module, we’ll dive deep into just some of the virtualization technologies that are exciting today. There wouldn’t be time to go into all of them, but I think you’ll be impressed with just how virtualization really helped power cloud.
48. vCPUs, vGPUs, and Memory
Virtualization really did give rise to cloud technology. I mean, it is one of those key ingredients that made the cloud revolution exactly the huge success that it was in our industry. Let’s talk about how common it is for us to be apportioning and subscribing to virtual CPUs, virtual GPUs, and virtual memory in the cloud.
All right, so here we are in the Google Cloud platform, and we are gonna see just how similar this is, of course, to the other public clouds that we have been examining. And let’s blow this font up so we can see it. Boy oh boy, that default font was very small inside of GCP, but notice the similarities to the other cloud platforms that we have been looking at. And it really is remarkable that they make it so easy for us to tap into these things like compute inside of the cloud. So here in Google Cloud platform, let me show you their All Products view.
Notice they don’t call them services, they call them products, interesting. But notice they do just what the other cloud providers do, we can see a Management category, a Compute category, a Storage category, and notice under the Compute, we have the Compute Engine.
Now, what’s interesting about Google Cloud, I’ve obviously been using this account that I’m demonstrating GCP with here for quite some time, but when you first come into the Compute Engine, you will have to enable it, and I really like that. I really think that’s such a neat touch and it is very distinct from AWS and Azure, where Google’s saying, ‘Look, you might not wanna use our cloud for compute services whatsoever, so don’t even enable them,’ right? So they’re just sitting there dormant, and obviously, that’s a reduced attack surface against your cloud, and all that kinda stuff, so that’s a really neat touch. But I have already enabled the Cloud Compute Engine and the appropriate APIs, so this management console will work, and I could even develop my own tools and scripts and all those would work, and once you’re fully enabled like that, you come in here, and sure enough, you are going to create things like virtual machines, and these are going to be built against differing amounts of virtual CPU, virtual memory, virtual graphics compute engines, if you need it, and notice, I love what Google Cloud platform does, just like AWS does, and Azure does now too, they all do this now that I think about it, but notice up in the top right, I love how bold they did it. Google Cloud platform is saying, ‘Look, this costs about 3 cents an hour.’ So just understand if you run a virtual machine with these settings that you have selected in here, we’re gonna be charging you about $25 a month for this privilege.
Now remember, again, that is if you left it on all the time. So, Google Cloud platform is gonna be just like AWS, in that if we stop the instance, we’re not gonna be charged for it, so that’s great. Now notice, this is what we wanted to focus on, look at this, this $25 a month virtual machine inside of Google Cloud platform, it is, given the fact that it’s one virtual CPU, and it’s four gigs of memory, if I go in here and say, ‘Whoa, wait a minute here, we need to really be doing some pretty intense calculation work here.’ In other words, maybe this machine is gonna be rendering big graphic files for us or something, and I need to step it up to four virtual CPUs with 16 gig, notice what that’s gonna do to our price.
Yeah, now we’re talking 100 bucks a month, yep, $100 a month. So, notice, as we manipulate those resources, we are gonna see differing prices, and, yes, I know you want me to do it, so let’s do it, let’s do a 32 virtual CPU system in the cloud with 32 gigs of memory. Now, before I click this and we look at the price, I just need to impress upon you that when you talk virtual memory and virtual CPUs and things, you are talking about very-very efficient ways to run an operating system. I’m gonna give away how old I am right now, unfortunately, but I have to tell you, I will never forget just my amazement when I saw, for the first time, Windows, and it was probably Windows 95, yeah, that’s how old the story is. When I saw Windows 95 running in a virtualized environment, and I was like, ‘Oh my goodness, this is performing incredibly well. How much memory does this have?’ And that’s when I was blown away, because the person I was working with said, ‘It’s one gig.’ And I thought for sure back then, it had at least two gig or four gig, which was a lot of memory way back then. So it’s amazing how you can give something just four gig of virtualized memory, and it’ll seem like a 16 gig system with its performance. All right, with that kinda side story out of the way, let’s go ahead and check this out.
It would be 32 virtual CPUs and 32 gigs of memory. Look, this is gonna be a very high cost. Eh, wow. Not at all as bad as I thought, right? $578 a month. Remember now, that would be leaving it running all the time, and we certainly wouldn’t be doing that. Typically, with a machine like this, this would probably be more special purpose. Fire it up, and then kind of forget about it. Not forget about it, rather, fire it up, use it, and then shut it down, and forget about it while it’s shut down.
Notice they were gonna give us a 10 gig boot disc by the way, for this device. Now, understand that that was just with the E-series here in Google Cloud platform. Now, what if I needed some graphics processing going on?
So, boy oh boy, you can get into some really intense configurations here. Notice they have these new Skylake-supported devices. So now look at this, if we come down here, we’re gonna see all kinds of configuration in this family. And notice that’s what’s happening, you are getting different configuration families based on the series that you’ve chosen. So my gosh, look at this high memory one, 96 virtual CPUs and 624 gigs of memory. My goodness, think of all of the virtual machines that you could run on that system. So here, you’re building a virtual machine to host all these virtualization, maybe containers or whatever. I mean, my goodness, it’s just an incredibly high RAM-outfitted machine, and now we’re getting up into the 3,000 per month range.
Now, as I said, there are going to be also virtualized graphics processing type systems that you can build. And so notice up at the top here, we have these tabs for the various machine families. Boy, I think we’re about to get into some big money now. So imagine this, you come in here and you say, ‘I need some intense graphics processing units of your GPU systems,’ right? ‘That I’m gonna be doing intense graphics processing with.’ And actually, this guy right here I think is the P4, I think. Yeah, and they won’t even give us that many GPUs. Yeah, I think we’re gonna get into some, no, no, that’s inexpensive. All right, well, anyways, you get the idea. Right at your fingertips, there’s 5,000 a month, woo, we’re getting pricey now. Right at your fingertips is the ability to go in and see the types of virtualized resources that you have access to right in the cloud. And yeah, don’t click Build, right? I would never come down here and click Create, because well, okay, look, it’s only $13 an hour. So, yes, I would do this, because it’s not like we’re gonna get a bill for 10,000 U.S. when we click Create. No, if I create this and then immediately destroy it, it’s gonna cost about 90 cents. So it’s fun to come in here and play and see the various options that are available.
And look at this, you can even come in and drill in to the specific CPU platform that you need. So, check this out in any of the big public clouds and you’re gonna see a similar paradigm. When you go to spin up that compute resource in the cloud, you’re gonna be placing it on a type of a machine and that machine type is gonna dictate how much RAM you get, how much virtual CPU you get, whether or not you’re gonna have GPUs in the system or not, all right there at your fingertips. Thanks for watching.
49. Other Virtualization Topics
Let’s go to the white board and let’s talk quickly about some other interesting virtualization topics that we need you to be aware of as a Cloud+ candidate.
So, when you think about the incredible virtualization capabilities today, when it comes to things like compute, we are amazed at the ever increasing clock speeds and capabilities. And sure enough, this hardware does exist in the cloud, and of course, it’s being virtualized to us. And the cloud really is keeping up. And in fact, the cloud is starting to offer hyper-converged type of environments for us to take advantage of. And this is a big private cloud concept right now as well. Cisco Systems is one of the ones that is in this market big time right now. They are selling these hybrid cloud solutions that are hyper-converged architectures. Now, what exactly does that mean? It means that the data storage and the networking components and the software that you need to make it all work are in one package. So now, keep in mind this is hardware that Cisco is selling for you to put in your own private data center, so that you can do total hybrid cloud where you have your own private cloud and then, you go up to AWS for things that you don’t want to do yourself privately, or maybe you’re going up to Azure, or Google Cloud platform, or IBM Cloud, or Rackspace, or insert any one of your favorite clouds out there today. There’s too many for me to name, right? But the big three obviously are AWS, Azure, and Google Cloud platform. And I just listed them in order of market penetration. The most successful financially of all three is Microsoft’s Azure, which is no big surprise, because they get so much revenue off of licensing, and all that, that’s often transferred into the cloud. So, you know, they own Windows, and Windows is being deployed all over the place now via the cloud, and they’re collecting revenue that way in their cloud. And yeah, a whole bunch of reasons. They lead in money, AWS leads in market penetration and I think both companies are just fine with that.
So, yeah, these types of virtualized, hyper-converged environments are very, very exciting. And just remember that, whatever advancements come along like, we were talking about GPUs, graphic processing units that allow you to get all this great horsepower. I mean, my laptop is outfitted with the GPU and the reason is, because I needed it for rendering videos. And also, I’m a huge flight simulator (FS) fan. And you know, I’m a private pilot, so I have a real appreciation for going into flight simulator and simulating instrument flight which I’m currently studying. So I need a graphics processing unit, like an NVIDIA card, and I think, it’s like a 2080 card in my laptop, I guess. And it’s the one of the most expensive components in the laptop. Well, now that, this technology is so popular, sure enough, we saw this in AWS, we saw it in Azure, you see it in Google Cloud platform. We see where they’re virtualizing now, the GPU. So any of these advancements that come along, when it comes to really super-fast processing, no surprise, we have the option of doing that inside of the cloud. Thanks so much for watching.
