2.7 Virtual Networking Part 7

00:00

So what is dynamic memory? So dynamic memory is a Microsoft hyper V feature first introduced in Windows Server 2008 are to believe it was

00:12

with the service park

00:14

on before we get started. So Lance has said, How do you know if Jen one or Jen to machine isn't determined by bios by you, Effy. So, Lance, Really good question. I wonder if maybe way didn't pick up on this on the previous one. So a Gen one machine will always use a BIOS environment on a gentle, always using UFO.

00:34

It's just the way it's being designed.

00:36

How do you know if a machine is jen one or Jen, too? When you go into the settings, you will notice that the hardware configuration, which when we went in and looked at the security settings, I covered it near the beginning. Effectively, we went in, and we saw that the machine I was on, which is a Gentoo machine,

00:55

had a scuzzy controller.

00:57

Now, if it was a Gen one machine, it would have had either an I D controller on door, a floppy disk drive,

01:06

a DVD drive on dhe, things like legacy network adapters and I d controllers and those kinds of things. So all the old kind of tech,

01:15

if you kind of keep down that line, that's the kind of Give away mark for whether a machine is genuine or Gentoo if he quickly are kind of in the Senate on which one. This is but yet genuine bios Environment Jen to you, Effy Environment, along with the respective kind of

01:34

generation of hardware changes that occur with it as well.

01:37

Really good question.

01:38

Yes, they're carrying on for this dynamic memory. So it came up in Server 2008 on it was expanded on in 7 2012 so this technology allows the hyper visor. So we discussed what hyper visor was in in the first section

01:56

is hyper visor

01:57

allows hyper visor toe handle ram consumption like memory consumption by host of the ends

02:05

in a flexible one, if possible. If it is possible, if there's not enough ram, it's not gonna do that at all. It's just going to stop. So, for example, wth e hyper V can dynamically ADM or memory to a virtual machine when the guest operating system needs it

02:23

and it can also reclaim it back. So if you run a program on the Vietnam and it goes, Oh, I've only got four. Gig of Ram. I need eight. It will go to the hardware to the hyper visor and go. Guys, I need more Ram adviser goes, Yep. I've got some of that idol punch over to you. You've you have now allocated this to please use it.

02:40

You do your work, and then that machine sits idle. You kind of log off. You go home for the night, and that machine is idle on the network. After a while, the hyper visor will go off and check in with it, and it'll go. Hey, you know that ram that I gave you? Are you done with it? And then the machine go? Yeah, I finished. You know, an hour ago that guys use has gone home.

03:00

Okay, I'm gonna take it back, and it will grab that and pull it back into this open pool,

03:05

ready to be used by other things. So it dynamically assigns memory to different machines

03:10

very often with machines on the network. You'll see that may be a server is only using two gig you wouldn't build generally a server with only two gig, and that's not doing much with it. But the second the server comes up and, you know, maybe you should start looking in throughout the day reading files, running applications on the server, that kind of thing.

03:30

It will start to request more, more. Resource is, and then towards the end of a lets go of the resource is,

03:35

and it allows this dynamic allocation of memory across the network, hence dynamic memory.

03:40

And it's a really, really flexible way off of being able to deal with. If you only have, say, 16 Gig of RAM, but you need servers that run up to 20 or 25 gig. It allows you to move the ram between the machines. How often does the host check? It is suitable.

03:59

You can set it by giving power shell commands to the host,

04:02

but generally the host itself by default, I believe checks every sort of 15 minutes or so on just kind of checks in going. Everything all right? Yeah. Need anymore? No, you're good. Okay. Can I take some back out? You're still using it. Okay, call. That's fine. So just checks in and out kind of gun yet. Okay, You're okay. You're okay.

04:21

It's kind of like having a manager in the middle on each virtual machine is like a stockman book.

04:27

Everything you need to know. Everything you need, everything you need. Yeah. Okay, cool. That's kind of how it sets up. So good question, Eric. Certainly. So, yeah, it is a flexible solution to memory management on dhe, you can add, As I said, you had more ground to a virtual machine when it's needed on, then reclaim it when it's no. So you pull it straight back.

04:44

The key advantage to this is memory is cheap. It has been cheap for a while.

04:49

However, it does deal with unlimited budget. So you might not. Maybe you've used up your annual budget for buying more memory or if you're on a home machine working. You're like, Well, I don't have I don't have another £200 to spend on more memory, so I'm just gonna deal with it in a dynamic fashion with my cluster.

05:10

It's also good for putting old service on the network. The reason that's important is we do cover it later on. There is an architecture built in that allows physical servers to share their memory, which is a really cool idea. So if you've got a server with 64 gig of round

05:28

and you go, Oh, my God. I need to run 70 gig worth of programs. Now

05:32

you can add another server in on a LL. That service job is if you want to run, just a old server is to do processing for that main server that you have. That 64 gig real powerhouse server on this server might only have eight gig, but it's enough to push you to that 70 year limit

05:50

on the main server will just send over

05:54

requests to the other server to deal with. It's a little bit slower, but you still have access to that memory, and then it will send it back again across the network. So real smart little system there.

06:04

So yeah, even though Ram is not much of an issue these days for pricing, RAM modules are noticeably dropping over the years in terms of price. Sometimes you can find this technology really handy, eso like if you're operating with a limited budget using old servers.

06:24

In addition to that,

06:26

this technology works perfectly with a like a great number of the EMS. Keeping like over utilization under control allows you to forget about wasting Resource is because it's all managed for you effectively, and it provides a better visibility towards your system.

06:43

On it can be especially really especially helpful

06:46

when you have a lot of idol or low loaded V EMS. Like if people remote into a machine to do their work during the day and then they go home. All that resource just comes back online. And maybe you can use that resource to run intensive update programs over night when there's no one around because they don't need that round

07:04

that zone. Really nice environment.

07:06

Eso yet on also it's quite handy. I didn't mention this one when there's faulty ram that's in the process of being replaced. It just means that you don't have to panic if you have to remove half the ram in the server and then you go, Oh my God, these machines are gonna work properly. They'll just pull it from another server effectively.