WEBVTT

0:00:03.260000 --> 0:00:08.020000
 Hello and welcome to this video titled
 an overview of hypervisors.

0:00:08.020000 --> 0:00:13.220000
 In this video I'm going to introduce
 you to what hypervisors are and no

0:00:13.220000 --> 0:00:15.120000
 this is not about your supervisor.


0:00:15.120000 --> 0:00:16.880000
 This is about a hypervisor.

0:00:16.880000 --> 0:00:21.400000
 We'll talk about type 1 hypervisors,
 type 2 hypervisors and I'll show

0:00:21.400000 --> 0:00:25.240000
 you some examples of both.

0:00:25.240000 --> 0:00:28.740000
 All right, so what
 is a hypervisor?

0:00:28.740000 --> 0:00:33.600000
 Well, in a non-virtualized environment,
 just take your typical laptop,

0:00:33.600000 --> 0:00:36.300000
 your typical server,
 your typical PC.

0:00:36.300000 --> 0:00:41.080000
 Your operating system, whatever it is,
 Windows operating system, Mac operating

0:00:41.080000 --> 0:00:45.740000
 system has exclusive access
 to the physical hardware.

0:00:45.740000 --> 0:00:49.680000
 So your operating system is what tells
 the CPU, hey, I need you to perform

0:00:49.680000 --> 0:00:50.700000
 a function for me.

0:00:50.700000 --> 0:00:54.640000
 Your operating system says, hey, RAM,
 here's some information I need or

0:00:54.640000 --> 0:00:56.720000
 I need to pull some information
 from you.

0:00:56.720000 --> 0:00:59.460000
 Your operating system has
 access to all of that.

0:00:59.460000 --> 0:01:05.060000
 In a virtualized environment, there
 can be multiple operating systems

0:01:05.060000 --> 0:01:07.280000
 on a single physical server.

0:01:07.280000 --> 0:01:13.440000
 You can have a physical server that's
 got Linux, Ubuntu, Microsoft, Cisco,

0:01:13.440000 --> 0:01:15.760000
 iOS, all sorts of different
 operating systems.

0:01:15.760000 --> 0:01:20.680000
 And the hypervisor is yet another operating
 system, but the hypervisor

0:01:20.680000 --> 0:01:25.980000
 is an operating system that sits between
 the hardware and these various

0:01:25.980000 --> 0:01:28.220000
 guest operating systems.

0:01:28.220000 --> 0:01:34.060000
 So we would say that the hypervisor
 is hosting the virtual machines.

0:01:34.060000 --> 0:01:38.340000
 That's why we call the hypervisor the
 host operating systems and the various

0:01:38.340000 --> 0:01:43.260000
 virtual machines, whatever they
 are, Windows, Cisco, iOS, Linux.

0:01:43.260000 --> 0:01:47.220000
 Each one of those individual operating
 systems is what we call a guest

0:01:47.220000 --> 0:01:48.960000
 operating system.

0:01:48.960000 --> 0:01:55.060000
 The hypervisor has access to the physical
 hardware of the host and provides

0:01:55.060000 --> 0:02:00.280000
 access to a set of emulated hardware
 for each virtual machine.

0:02:00.280000 --> 0:02:04.740000
 And it does this in a way that's transparent
 to the guest operating system.

0:02:04.740000 --> 0:02:09.920000
 So for example, if I have a virtual machine
 that is Windows, for example,

0:02:09.920000 --> 0:02:13.320000
 let's say I have a virtual machine
 that's a Windows 10 desktop virtual

0:02:13.320000 --> 0:02:18.420000
 machine. That virtual machine will think
 when it's performing a call to

0:02:18.420000 --> 0:02:24.280000
 the CPU or performing a call to memory,
 it will think Windows on that

0:02:24.280000 --> 0:02:29.100000
 virtual machine Windows will think
 it has access direct access to the

0:02:29.100000 --> 0:02:30.500000
 CPU, to the RAM.

0:02:30.500000 --> 0:02:33.220000
 But in reality, it doesn't.

0:02:33.220000 --> 0:02:38.800000
 Those requests it's making actually go
 to the hypervisor, another operating

0:02:38.800000 --> 0:02:43.480000
 system that's controlling and manipulating
 and scheduling everything.

0:02:43.480000 --> 0:02:47.540000
 And then that hypervisor will have to
 say, okay, hold on from microsecond.

0:02:47.540000 --> 0:02:51.200000
 Actually, won't tell Microsoft, hold
 on, I'm going to take your request.

0:02:51.200000 --> 0:02:56.260000
 But right now, I've got another VM over
 here that I sent his request to

0:02:56.260000 --> 0:03:00.820000
 the CPU in a nanosecond when
 the CPU is done with him.

0:03:00.820000 --> 0:03:04.420000
 Now I'll take your request
 and send it to the CPU.

0:03:04.420000 --> 0:03:08.700000
 So each virtual machine has no idea that
 there are other virtual machines

0:03:08.700000 --> 0:03:13.880000
 that are sharing the same physical resources,
 the hypervisors and control

0:03:13.880000 --> 0:03:19.060000
 of all of that. So a definition we
 can see here is that it's a process

0:03:19.060000 --> 0:03:23.120000
 that separates the computer's operating
 system and applications from the

0:03:23.120000 --> 0:03:25.540000
 underlying physical hardware.

0:03:25.540000 --> 0:03:30.980000
 Hypervisors are typically implemented
 as software and it maximizes the

0:03:30.980000 --> 0:03:35.080000
 efficient use of the hardware
 in the actual host system.

0:03:35.080000 --> 0:03:38.200000
 And it keeps the virtual
 machines separated.

0:03:38.200000 --> 0:03:40.200000
 So this is what it looks like.

0:03:40.200000 --> 0:03:45.580000
 So here we have a couple of virtual
 machines, guest OS, maybe virtual

0:03:45.580000 --> 0:03:50.920000
 machine one is Windows, maybe virtual
 machine number two is a Cisco iOS

0:03:50.920000 --> 0:03:57.220000
 router. And then we also have our base
 operating system, which is our

0:03:57.220000 --> 0:04:01.900000
 hypervisor. Now, our guest operating
 system is going to make some calls

0:04:01.900000 --> 0:04:03.780000
 for physical resources.

0:04:03.780000 --> 0:04:05.920000
 Maybe one of those calls
 is for the CPU.

0:04:05.920000 --> 0:04:10.780000
 Maybe another call is to direct something
 to the virtual screen, whatever

0:04:10.780000 --> 0:04:15.860000
 it might be. The second VM could also
 be making some calls at exactly

0:04:15.860000 --> 0:04:21.180000
 the same time. Well, clearly if those
 calls are for the CPU at exactly

0:04:21.180000 --> 0:04:26.940000
 the same time, the hypervisor can't
 connect to the physical CPU and send

0:04:26.940000 --> 0:04:29.340000
 it multiple instructions
 all at once.

0:04:29.340000 --> 0:04:31.680000
 So it has to schedule these.

0:04:31.680000 --> 0:04:35.360000
 So the hypervisor is actually what
 connects to the physical resources

0:04:35.360000 --> 0:04:41.820000
 and makes those resources
 available to the VMs.

0:04:41.820000 --> 0:04:46.220000
 The other thing I'll say about this
 is that each virtual machine only

0:04:46.220000 --> 0:04:52.800000
 sees the amount of CPU, RAM and hard disk
 drive space that has been allocated

0:04:52.800000 --> 0:04:54.860000
 to that virtual machine.

0:04:54.860000 --> 0:04:58.700000
 So for example, when you're creating
 a virtual machine in the first place,

0:04:58.700000 --> 0:05:00.420000
 there's a lot of steps
 that are involved.

0:05:00.420000 --> 0:05:03.940000
 For example, when you create a virtual
 machine, it's done the hypervisor.

0:05:03.940000 --> 0:05:08.360000
 So the hypervisor is what gives you the
 GUI or the command line to create

0:05:08.360000 --> 0:05:10.660000
 virtual machines.

0:05:10.660000 --> 0:05:13.000000
 And so when you're creating virtual machines,
 you'll have to specify things

0:05:13.000000 --> 0:05:16.900000
 like, okay, what's going to be the operating
 system of the guest of the

0:05:16.900000 --> 0:05:21.260000
 virtual machine will be Windows, will
 be Linux, will be Cisco iOS, what

0:05:21.260000 --> 0:05:25.540000
 will be. Then you'll also have to specify
 things like, okay, how many

0:05:25.540000 --> 0:05:29.240000
 CPUs will be allocated to
 this virtual machine?

0:05:29.240000 --> 0:05:34.320000
 Because maybe in your physical box,
 you've got, you know, eight sockets

0:05:34.320000 --> 0:05:38.800000
 and each socket has four cores in
 or capable of multi-threading.

0:05:38.800000 --> 0:05:43.560000
 So you've got some variations here as
 far as how many CPUs and how many

0:05:43.560000 --> 0:05:46.640000
 virtual CPUs are available
 to each virtual machine.

0:05:46.640000 --> 0:05:47.740000
 You'll have to plug that in.

0:05:47.740000 --> 0:05:49.760000
 How many CPUs do you need?

0:05:49.760000 --> 0:05:52.980000
 How much memory do you need each virtual
 machine to have as far as like

0:05:52.980000 --> 0:05:54.280000
 RAM is concerned?

0:05:54.280000 --> 0:05:56.600000
 You'll plug that in
 as a number as well.

0:05:56.600000 --> 0:06:01.100000
 Now once you do that, let's say for example
 that we tell the virtual machine,

0:06:01.100000 --> 0:06:08.240000
 hey, you've got a hard disk space
 of two gigabits, let's just say.

0:06:08.240000 --> 0:06:13.240000
 Well, even though the physical host
 might have a terabit hard drive, the

0:06:13.240000 --> 0:06:14.500000
 virtual machine doesn't know that.


0:06:14.500000 --> 0:06:17.820000
 You told the virtual machine, you get
 two gigabits, so that's what he

0:06:17.820000 --> 0:06:21.600000
 sees. That's how much memory
 he thinks is available.

0:06:21.600000 --> 0:06:25.920000
 He does not have visibility to
 the actual physical resources.

0:06:25.920000 --> 0:06:28.380000
 Only the hypervisor can see that.

0:06:28.380000 --> 0:06:32.600000
 Now there's two different
 types of hypervisors.

0:06:32.600000 --> 0:06:35.700000
 The first one I want to talk about here
 is what's called a Type 1 hypervisor.

0:06:35.700000 --> 0:06:39.540000
 This is also called a native
 or bare metal hypervisor.

0:06:39.540000 --> 0:06:44.140000
 This runs directly on
 the host's hardware.

0:06:44.140000 --> 0:06:49.680000
 For example, this will be a situation
 where you buy a server without an

0:06:49.680000 --> 0:06:50.580000
 operating system.

0:06:50.580000 --> 0:06:54.120000
 You would go and you'd say, okay, let
 me buy a server with this amount

0:06:54.120000 --> 0:06:58.440000
 of CPU, this amount of RAM,
 maybe this graphics card.

0:06:58.440000 --> 0:07:02.660000
 Then just like you would normally load
 Linux as an operating system or

0:07:02.660000 --> 0:07:07.300000
 maybe you load Windows Server as an
 operating system, instead you would

0:07:07.300000 --> 0:07:12.320000
 buy a Type 1 hypervisor and load
 that as the operating system.

0:07:12.320000 --> 0:07:16.720000
 This provides a direct interface between
 the virtual machines and the

0:07:16.720000 --> 0:07:17.900000
 hardware resources.

0:07:17.900000 --> 0:07:21.360000
 Some examples of this
 are Citrix Zen.

0:07:21.360000 --> 0:07:24.600000
 This is actually an open
 source Type 1 hypervisor.

0:07:24.600000 --> 0:07:27.380000
 It doesn't cost you anything
 to use this.

0:07:27.380000 --> 0:07:32.040000
 Oracle has one, actually a couple.


0:07:32.040000 --> 0:07:36.680000
 Microsoft has their Hyper-V and
 a very popular one is VMware.

0:07:36.680000 --> 0:07:40.400000
 They have something
 called ESX and ESXI.

0:07:40.400000 --> 0:07:43.880000
 There's also KVM which
 is built into Linux.

0:07:43.880000 --> 0:07:44.860000
 Here's an example.

0:07:44.860000 --> 0:07:47.660000
 Here's a screenshot of
 a Type 1 hypervisor.

0:07:47.660000 --> 0:07:52.520000
 This happens to be the screenshot
 of VMware's ESXI.

0:07:52.520000 --> 0:07:56.500000
 Once you load ESXI,
 this is the GUI.

0:07:56.500000 --> 0:07:57.160000
 It would give you.

0:07:57.160000 --> 0:08:00.360000
 From here, you can build
 your virtual machines.

0:08:00.360000 --> 0:08:01.940000
 You can see how they're doing.

0:08:01.940000 --> 0:08:03.100000
 You can manipulate them.

0:08:03.100000 --> 0:08:05.060000
 You can turn them on
 or turn them off.

0:08:05.060000 --> 0:08:10.320000
 This is the base operating system that
 you load onto the physical hardware.

0:08:10.320000 --> 0:08:14.000000
 This is a Type 1 hypervisor.

0:08:14.000000 --> 0:08:18.800000
 You might say to yourself, but Keith,
 I need to build a virtual machine

0:08:18.800000 --> 0:08:25.240000
 on my laptop. I don't have the resources
 to go out and buy a physical

0:08:25.240000 --> 0:08:29.060000
 server. I wish I could, but all I've
 got available to me is a laptop.

0:08:29.060000 --> 0:08:32.840000
 This laptop is already
 running some software.

0:08:32.840000 --> 0:08:36.840000
 It's already running Windows
 or it's already running Mac.

0:08:36.840000 --> 0:08:39.520000
 For example, let's say Mac.

0:08:39.520000 --> 0:08:41.640000
 It's already running the
 Mac operating system.

0:08:41.640000 --> 0:08:45.880000
 Yet, I want to load an application on here
 that's a Windows-based application

0:08:45.880000 --> 0:08:48.140000
 that doesn't run on Mac.

0:08:48.140000 --> 0:08:49.940000
 How am I supposed to do that?

0:08:49.940000 --> 0:08:54.120000
 That's where you would use something
 called a Type 2 hypervisor.

0:08:54.120000 --> 0:08:59.240000
 This is what's called a hosted hypervisor
 because the hypervisor itself

0:08:59.240000 --> 0:09:05.260000
 does not have direct access to the physical
 memory, CPU power, all that

0:09:05.260000 --> 0:09:09.800000
 stuff. There's already another operating
 system, your base operating system

0:09:09.800000 --> 0:09:11.380000
 that's already running.

0:09:11.380000 --> 0:09:17.300000
 A hosted hypervisor or a Type 2 hypervisor
 runs on top of conventional

0:09:17.300000 --> 0:09:20.840000
 operating systems as
 another application.

0:09:20.840000 --> 0:09:23.400000
 It still provides you
 this separation.

0:09:23.400000 --> 0:09:29.440000
 Once you have a Type 2 hypervisor, for
 example, VMware Workstation, VMware

0:09:29.440000 --> 0:09:34.680000
 Player, you can see these here are some
 examples of Type 2 hypervisors.

0:09:34.680000 --> 0:09:39.600000
 Once you load this, let's say you
 use VirtualBox as an example.

0:09:39.600000 --> 0:09:43.060000
 Here's an example of what
 VirtualBox looks like.

0:09:43.060000 --> 0:09:49.000000
 I'll show it to you right here.

0:09:49.000000 --> 0:09:53.520000
 I'm using Mac. This presentation
 is being done on a MacBook.

0:09:53.520000 --> 0:09:57.000000
 You've got your typical
 Mac background here.

0:09:57.000000 --> 0:10:00.620000
 I downloaded Oracle's VirtualBox.

0:10:00.620000 --> 0:10:04.120000
 The great thing about VirtualBox
 is that it's free.

0:10:04.120000 --> 0:10:06.260000
 This is a hypervisor.

0:10:06.260000 --> 0:10:09.760000
 It was downloaded as another app, so
 it's running as just another app

0:10:09.760000 --> 0:10:12.560000
 on top of my Mac operating system.


0:10:12.560000 --> 0:10:14.400000
 This is a special app.

0:10:14.400000 --> 0:10:17.920000
 This is a Type 2 hypervisor because
 once I've got Oracle's VirtualBox

0:10:17.920000 --> 0:10:21.920000
 running, now I can create
 VMs inside of it.

0:10:21.920000 --> 0:10:25.500000
 For example, you can see here
 I've got some Windows VMs.

0:10:25.500000 --> 0:10:30.120000
 If I start up this Windows 7 right
 here, this will actually start up.

0:10:30.120000 --> 0:10:31.280000
 I'll actually go ahead and do it.

0:10:31.280000 --> 0:10:34.340000
 I don't know how long it's going to take,
 but let's just see what happens.

0:10:34.340000 --> 0:10:41.760000
 Once I start up my Windows 7 VM, then
 I'll actually have, right here,

0:10:41.760000 --> 0:10:47.540000
 it's starting up right now, a Windows
 7 laptop you could think of running

0:10:47.540000 --> 0:10:51.360000
 inside of my MacBook.

0:10:51.360000 --> 0:10:52.860000
 This is a virtual machine.

0:10:52.860000 --> 0:10:54.280000
 Good, it works. There we go.

0:10:54.280000 --> 0:11:01.680000
 Now I've got Windows 7 and once Windows
 7 boots up, I could run my Windows

0:11:01.680000 --> 0:11:04.420000
 applications inside of here.

0:11:04.420000 --> 0:11:10.220000
 Now there's pros and cons to both
 Type 1s and Type 2 hypervisors.

0:11:10.220000 --> 0:11:12.760000
 Let's start with the Type 2s since
 that's what we're looking at right

0:11:12.760000 --> 0:11:18.340000
 here. The Type 2 hypervisor is nice
 because, number one, many of them

0:11:18.340000 --> 0:11:22.900000
 are free. VirtualBox is free.

0:11:22.900000 --> 0:11:29.720000
 You've still got your base operating
 system like my Mac operating system

0:11:29.720000 --> 0:11:32.900000
 that I can do all sorts of things with
 that, even if I choose not to use

0:11:32.900000 --> 0:11:34.300000
 VirtualBox at all.

0:11:34.300000 --> 0:11:41.620000
 The downside to a Type 2 hypervisor
 is you've got now one more layer of

0:11:41.620000 --> 0:11:46.920000
 abstraction. What I mean by that is that
 whatever application I have running

0:11:46.920000 --> 0:11:51.880000
 here inside of this Windows
 virtual machine.

0:11:51.880000 --> 0:11:58.100000
 Let's say I start up Google
 Chrome in here.

0:11:58.100000 --> 0:12:00.680000
 When Google Chrome starts
 up, what's it going to do?

0:12:00.680000 --> 0:12:02.860000
 He's going to ask for
 some resources.

0:12:02.860000 --> 0:12:06.120000
 He's going to ask for some like
 RAM memory for an example.

0:12:06.120000 --> 0:12:07.100000
 What's going to happen?

0:12:07.100000 --> 0:12:11.840000
 Well because this is a Type 2 hypervisor,
 when he requests RAM memory,

0:12:11.840000 --> 0:12:16.580000
 that Type 2 hypervisor VirtualBox is
 going to have to make the request.

0:12:16.580000 --> 0:12:18.640000
 Where's that request going to hit?


0:12:18.640000 --> 0:12:21.260000
 It's going to hit my
 Mac operating system.

0:12:21.260000 --> 0:12:25.160000
 Then my Mac operating system in turn
 will have to access the physical

0:12:25.160000 --> 0:12:28.120000
 RAM to give it back to VirtualBox.


0:12:28.120000 --> 0:12:31.120000
 We've got several layers
 of abstraction here.

0:12:31.120000 --> 0:12:35.480000
 It can cause your virtual machines to
 run a little bit slower to not quite

0:12:35.480000 --> 0:12:37.720000
 as be as efficient.

0:12:37.720000 --> 0:12:42.380000
 When you're talking about virtual machines
 that need a lot of CPU, the

0:12:42.380000 --> 0:12:48.400000
 need to access memory very quickly
 to read and write to memory, a Type

0:12:48.400000 --> 0:12:52.880000
 2 hypervisor might not
 be the best scenario.

0:12:52.880000 --> 0:12:58.580000
 Instead, a Type 1 hypervisor, which
 we saw an example of that right here

0:12:58.580000 --> 0:13:05.880000
 in VMware's ESXi, this is more efficient
 because a Type 1 hypervisor has

0:13:05.880000 --> 0:13:09.080000
 direct access to your
 physical resources.

0:13:09.080000 --> 0:13:13.180000
 There's no other operating system underneath
 this that has to pass through

0:13:13.180000 --> 0:13:18.160000
 first. Type 1 hypervisor will
 get you better performance.

0:13:18.160000 --> 0:13:20.220000
 It'll give you better scalability.


0:13:20.220000 --> 0:13:24.400000
 The downside though is a Type 1 hypervisor,
 that is the operating system

0:13:24.400000 --> 0:13:28.480000
 that's installed right on top
 of the physical hardware.

0:13:28.480000 --> 0:13:32.000000
 You're not going to be able to take
 your MacBook or your typical Windows

0:13:32.000000 --> 0:13:36.880000
 laptop, strip off your Mac operating system,
 strip off your Windows operating

0:13:36.880000 --> 0:13:41.640000
 system and install VMware's
 ESXi on top of that instead.

0:13:41.640000 --> 0:13:43.160000
 You won't be able to do that.

0:13:43.160000 --> 0:13:46.380000
 So with a Type 1 hypervisor, typically
 that means you've got to go out

0:13:46.380000 --> 0:13:47.840000
 and buy yourself some hardware.

0:13:47.840000 --> 0:13:51.820000
 You have to buy yourself a server with
 a lot of memory, a lot of CPU,

0:13:51.820000 --> 0:13:56.640000
 install the Type 1 hypervisor on top of
 it, and now whatever VMs you create

0:13:56.640000 --> 0:14:01.260000
 inside of there will be a lot more efficient,
 a lot faster than the same

0:14:01.260000 --> 0:14:05.600000
 VM if it was running inside
 of a Type 2 hypervisor.

0:14:05.600000 --> 0:14:12.900000
 So that concludes this section on hypervisors
 and the differences between

0:14:12.900000 --> 0:14:14.700000
 Type 1s and Type 2s.

0:14:14.700000 --> 0:14:16.160000
 I hope you found this enjoyable.