1
00:00:00,940 --> 00:00:04,940
Hyper‑V Discrete Device Assignment is a Hyper‑V feature

2
00:00:04,940 --> 00:00:09,030
that allows you to pass an entire PCIe hardware device,

3
00:00:09,030 --> 00:00:15,020
examples would be graphics cards, NVMe storage devices, into the VM.

4
00:00:15,020 --> 00:00:16,540
And when I say pass them in,

5
00:00:16,540 --> 00:00:19,950
we're not talking about simply linking like we did in the

6
00:00:19,950 --> 00:00:23,870
previous module, where the enhanced session configuration allows

7
00:00:23,870 --> 00:00:26,260
the VM to get to the host's hard drive.

8
00:00:26,260 --> 00:00:26,550
No,

9
00:00:26,550 --> 00:00:29,690
this is where you're actually giving a piece of hardware

10
00:00:29,690 --> 00:00:32,360
on the host exclusively to the VM.

11
00:00:32,360 --> 00:00:36,890
We configure this using PowerShell. And we just need to know the very

12
00:00:36,890 --> 00:00:40,550
high‑level bird's eye perspective for our exam success.

13
00:00:40,550 --> 00:00:45,140
Basically, you have to configure the VM to support Discrete Device Assignment,

14
00:00:45,140 --> 00:00:49,560
you'll have to work on the host to dismount the device, and then assign the

15
00:00:49,560 --> 00:00:55,880
device to the guest VM. Hyper‑V central processor unit, or CPU, groups allow

16
00:00:55,880 --> 00:01:01,710
an administrator to get better control over the host CPU and the guest VM

17
00:01:01,710 --> 00:01:07,320
access to host CPU. This is analogous to what we did in the previous module

18
00:01:07,320 --> 00:01:12,430
with dynamic memory and different things you can do like memory weighting to

19
00:01:12,430 --> 00:01:18,090
allocate host RAM to your VMs in the way that you want to or that you need to

20
00:01:18,090 --> 00:01:18,900
for your business.

21
00:01:18,900 --> 00:01:23,160
Well, here we have a similar capability, a Quality of Service capability, for

22
00:01:23,160 --> 00:01:29,700
the host CPU. Now, we need to access an API called the Hyper‑V Host Compute

23
00:01:29,700 --> 00:01:34,930
Service, or HCS, and as of this recording in early 2022, and even with the

24
00:01:34,930 --> 00:01:39,590
latest version of Windows Server, Windows server 2022, we don't have a way of

25
00:01:39,590 --> 00:01:45,620
accessing the HCS outside of a separate executable. Microsoft created an

26
00:01:45,620 --> 00:01:50,930
executable called CPU groups that's essentially an abstraction over the Host

27
00:01:50,930 --> 00:01:52,020
Compute Service.

28
00:01:52,020 --> 00:01:52,410
Again,

29
00:01:52,410 --> 00:01:56,330
we don't need to get deep into the weeds here, you just need to identify the

30
00:01:56,330 --> 00:02:01,040
capability if you're asked about it on exam AZ‑800, and basically what you're

31
00:02:01,040 --> 00:02:08,210
doing is using that CPU group's executable to divide logical VM resource groups

32
00:02:08,210 --> 00:02:13,760
into CPU service tiers. I know that the exam AZ‑800 objectives reference

33
00:02:13,760 --> 00:02:18,610
resource groups. There's not a similar concept to Azure resource groups in

34
00:02:18,610 --> 00:02:22,530
Hyper‑V, so I believe what they're talking about, what Microsoft is talking

35
00:02:22,530 --> 00:02:29,610
about there is strictly CPU groups. So we divide our logical VMs into these CPU

36
00:02:29,610 --> 00:02:35,530
service tiers, where you may have one or more VMs that need more CPU power,

37
00:02:35,530 --> 00:02:40,770
just simple as that, and more reserved CPU power than others, and you may also

38
00:02:40,770 --> 00:02:46,450
want to use CPU groups to constrain CPU on VMs to make sure that they don't go

39
00:02:46,450 --> 00:02:51,030
out of bounds. That make sense? I do have just some example usages. I doubt very

40
00:02:51,030 --> 00:02:55,870
seriously you'd see CPU groups syntax on your exam, so I'm giving you this

41
00:02:55,870 --> 00:03:01,310
basically for color. We could do CpuGroups GetCpuTopology just to see what your

42
00:03:01,310 --> 00:03:03,440
current host topology is.

43
00:03:03,440 --> 00:03:08,400
This example shows us creating a group that includes four logical processors.

44
00:03:08,400 --> 00:03:11,510
In other words, these are processor cores on the machine.

45
00:03:11,510 --> 00:03:16,390
Third example shows us how we can set a CPU cap to 50%.

46
00:03:16,390 --> 00:03:22,010
Now, notice that the range is 0 to 65536, so to do a 50% cap,

47
00:03:22,010 --> 00:03:24,070
you specify half that value.

48
00:03:24,070 --> 00:03:25,870
It's a 16‑bit value.

49
00:03:25,870 --> 00:03:30,030
Now, of course, you're probably thinking, whoa, and yeah,

50
00:03:30,030 --> 00:03:33,040
this is not user friendly in the least.

51
00:03:33,040 --> 00:03:36,360
Check the exercise files if you're thinking of actually implementing

52
00:03:36,360 --> 00:03:39,490
this. It's going to have a little learning curve to it, because it's

53
00:03:39,490 --> 00:03:42,570
about as anti‑intuitive as you can get.

54
00:03:42,570 --> 00:03:47,110
So you create your groups that allocate CPU cores from the host,

55
00:03:47,110 --> 00:03:51,030
you create your group properties that customize the behavior and

56
00:03:51,030 --> 00:03:57,000
the quotas of the groups, and then lastly, you bind VMs to the CPU group.