HPE Server Management with Compute Ops Management and iLO

Hi everyone. Uh, my name is Chris Powell. I'm in the HP Compute Technical Enablement team, and you saw my boss already present.

I'm joined here today by Andrew Leeky, and I'm also part of the compute technical enablement team. Do you have The same boss? I do have the same boss, yes.

So we work together closely on all of these things that you're about to see. All right. So we're gonna take you through more of a live demonstration of IO seven of compute ops management, uh, touch one view and the rest of the portfolio of management.

From a management perspective, we divide up the presenting duties. Is that correct? That's correct.

I'm gonna take over about halfway once Chris forgets what he's doing, because, you know, That could happen. Yeah. Yes, sir.

Alright, so I'll grab it then. Yeah, of course. Take the first part.

So, um, our, our systems in question here are, are on the fifth floor of this building. And, uh, we run it like a data center. It's not as formal as some data centers.

We're not using, uh, necessarily retina scans to let access in and out of the doors. Um, however, uh, we try to run it like a data center. We try to keep up with the inventory.

Uh, we're managing it, uh, from our GreenLake platform. And the reason I wanted to open with GreenLake is just to make sure that we're grounded in all of the other services that this platform provides besides the management of the compute, uh, systems themselves, right? So you're gonna see the data services piece, uh, the sustainability and insight center, which is really that AI story that we're talking about where we can, um, uh, watch the systems over time and deduce their trends and how much their, uh, electricity they're consuming, what, where you might save, uh, things from an operational perspective.

There was ops ramp, um, Chris Bradley talked about that, uh, briefly, uh, private cloud, the Aruba piece for networking. But why we're here today, let's talk about compute ops management. All right?

So let me go to compute ops management, and this is a real world thing, right? I, I want to do a demonstration of the IO seven of a brand new gen 12 server. Alright?

So I'm gonna do a demo ILO seven brand new Gen 12 server. And one of the ways, one of the old fashioned ways is of, of course, to have memorized all the IP addresses necessary to connect to wherever inside the environment. Uh, but that's the hard way, alright?

I have to recall all of that. I have to open up some kind of, um, inventory, something to know where stuff is. We're trying to create an intuitive platform that allows you to quickly identify stuff from an operations perspective so that you could just go to it and do things without having to be trained on, uh, the use case of the model.

Uh, remember the old days you compile code, you'd have to actually open the user's guide. Well, this is supposed to be intuitive. You're supposed to be able to just look at it and know what to do next type of thing, right?

So right away, you know, we're looking at, um, this geographic view. And the reason, again, we're doing this is, you know, as you're gonna deploy in these edge systems, wouldn't it be neat if you could see all your edges, see all your clusters of computers geographically dispersed in whatever area that you're in, so that you can then drill down on 'em, click through to them, and start the management process again, without having to look into an inventory spreadsheet. That's the idea, right?

So in this case, I'm trying to use the Gen 12 server, trying to find one so that I can do, uh, this demonstration. So I just go to the list of servers. Again, no user's guide.

It's supposed to be intuitively obvious. I'm gonna sort by model. All right?

And with a little bit of luck at the top of the list right here, sure enough, it's a couple of Gen 12 servers. So mission accomplished. I knew I had two gen 12 and this a hundred server plus I needed to go to those gen twelves.

Notice I'm clicking through with the authentication. So we've got it set to click through. And so from the, you know, GreenLake comm, I was able to enter into my data center.

And now as if I was sitting in front of that computer with a little laptop and a nick plugged into that server directly, I'm right there. Now, I, granted if, uh, the, um, service operation that I needed to do right this second required me to open the chassis and do something inside the computer, I would not be able to do that. I'm physically not there, right?

It's over there on the fifth floor in another part of the building, right? But everything else, cycle of power, figure out what's wrong with it. Um, turn it on and off and all those aspects of it I can do as if I was sitting in front of it.

And one of the first things that you've noticed, especially if you're familiar with our other baseboard management controller, the predecessor to this I oh five and six, um, is that, that that solution had grown over time. I've been here the entire time that we introduced ilo. You saw that chart, um, uh, earlier that Darren put up.

The, the different, many of us in the room lived through all of that. I remember the actual standup card back in the day. But anyway, the menus had grown over time to expand functionality.

And maybe the organization wasn't as clear as what you might do if you did a do over. Does that make sense? Because you're growing the thing, the ecosystem over time, like two decades.

So finally with this, we got to rethink it and say, why not lay this stuff out in a, in a manner in which people work. So why in the world are you on this device to begin with? I am gonna throw that out to y'all.

Why might you? What's the first use of this device? Onboarding.

Onboarding, getting the computer up and running, right? You know what's on the thing. Load it up.

Uh, put our support, our drivers and, uh, firmware support the rest of the stuff on the computer on break. All right? What's the next reason?

You're on the front of this computer broke. Something's broken. Something broke, right?

Break fix, right? This is a security warnings, but it's, you're in some kind of break fix situation. You have to figure out if it's operating system, hardware, firmware, what, whatever it might be.

Uh, and you're gonna want to deal with the host and work with that. Another one. Why would you be on that server?

Gonna run, you're on a roll, Tina. I'll give you a hint. Install software.

Yes. Yes. The initial deployment of the system, right?

Going into the virtual media mounting the media you might install and so on and so forth. Um, another piece, yeah, I'll help you out. Security.

Yeah, you'll be there doing some se security aspect of it, probably most likely at the beginning, but you're in here doing things, right? So the point is, again, and then you have the settings obviously for the ILO itself. So the point is that we try to organize this thing so that, again, it's intuitive doesn't need a user's guide.

You know what to do by looking at it. You click on the hyperlinks things happen, that kind of thing. Alright?

So with that, I'm gonna talk about security real quick 'cause this was two aspects that Cole brought out, okay? Um, and there's a number of different pieces to this obviously, right? So one of 'em is the encryption security.

So you gotta communicate to the space board management controller. You have to use some method to connect to it. And what we're gonna do is tighten up those connections to make sure that there's secure, alright?

So the first thing that we're gonna do is make them a little bit more secure than the predecessor. This, the, the predecessor five and six ISOs had a lot of security features in it, but this one goes a little bit further, alright? 3, right?

So the base, this I I'm sure they mentioned this, we're engineering security into the product by default. That's one of the engineering design goals. Is that the thing?

Be more secure. Alright? So right away connecting on the interface where that IO is we're gonna cut down all the ports, make sure there's only three ways in, and that you have to conform to these security pieces as well to successfully connect.

That will make it more secure, obviously, right? So that's not enough. There can be higher levels of security and for instance, we can move it up a notch to FIPs.

And we mentioned this kind of capability when we talked about the launch and the server, right? All of this, um, FIPs one 40 dash three level three, right? All this stuff.

So you're getting more of this capability as you move forward with making it a more secure connection to use the device. All right? Um, and you're encrypting stuff as you move through this.

So the action to make this happen, it's pretty easy. You gotta update security state, I'm sorry, you go to update security state, which is where I am. Oh, I have in sufficient privilege.

You see that? That was cool. I'm in on the wrong with the wrong user.

See that right there? So I can't change my security state forward Okay? That that's why I've use the other server.

Yeah. But that's the kind of thing. So, and I, I am in as a HPE user at GreenLake, that's been passed through down to this server and I have certain capabilities and one of the ones I don't have is full admin rights, right?

So it's good. That's how you lock this stuff up and keep people from flipping things. By the way, as you move forward in security levels, uh, you reset the ilo, if you go backwards, you have to completely factory reset it, which means you lose the passwords and have to get the toe tacky Question.

It's, this is the GreenLake interface. But if you've purchased a computer CapEx regularly still same thing, right? I can still use this or do I have to be an actual GreenLake Customer?

So let me, let me distill that a little bit. The astute observation was that looks a lot like GreenLake and yes sir, we're trying to make our interfaces on our management apps look the same. That is not at all GreenLake right now.

I'm actually down on a computer connected to its ILO and I could have got this through GreenLake. I, that's how I found the machine, but I could have gone straight to it if I knew the IP address, right? So I'm, I'm not, I'm kind of not GreenLake right now.

I'm on the machine. Well, you went through comm Yes, it's GreenLake to, uh, I picked one of the apps, calm and then calm in. Yeah.

So, yeah. Yes sir. It's through Calm technically Yeah.

Is something I don't need. GreenLake, I can have Calm, which n No, our solution stack is GreenLake, which hosts calm the storage stuff, the Aruba stuff, all that, and then calm. So you're gonna take that sub app and that's what, um, I'm not using the sub app right now.

Uh, Andrew's gonna come up and show it. I'm on the actual individual machine. The ILO itself.

Well, I'm just saying if I have a fleet, if I buy a Fleet CapEx right, I buy it. Yeah. Yeah.

Mm-hmm. Mm-hmm. I, but my control plane will still be as long as I have it under support the GreenLake.

Yes, sir. Um, the GreenLake starts it all. Alright.

Your machines will be individually e listed and calm. You'll be able to talk to 'em through automation, which means through the rest APIs in mass or one at a time. Like I'm doing, I'm just saying I don't have to have like one of the specific GreenLake solutions purchase through that vehicle.

I can get it through. Yeah. So financial question, I don't Yeah, yeah, yeah, yeah.

It's not a technical question. Um, right. And the financial question, go ahead.

The user account into GreenLake is free anybody, and you can get in the way you use stuff like, um, computer ops management or Aruba Central or, um, A storage Storage console, um, is licensing that management app. And then once you have the license, you apply the key IT to your account and then you have access to all of the SaaS apps that live within the GreenLake cloud. So Yeah.

Does that make sense? If I don't License the management app, I have to directly go to the machine. Yes, sir.

Yes, sir. That's correct. That's, That's the web-based io.

Yeah. Yeah, Yeah. So, and that's actually very similar to a lot of, uh, a lot of other systems where, you know, the cloud, the cloud connected part is the part you're paying for where the direct connect part is widely available.

And this has how been, how it is on proli forever. I mean, you've, yeah, We have tiered levels of capability. So you get some base support, like text base, um, consoles being able to, uh, turn the machine on and off, things like that.

But when you wanted to start mounting media, get, uh, a gooey, you licensed up the level is basically what happened. And so, so what we have is two aggregators of the IO information and it's either comm or one view. One view is on on site management tool.

I'm gonna show that in the end. And then comm is on, on the cloud, but they're basically IO aggregators in a sense, but showing you all the machines in the one view, you know what I mean? And you can also see the KVM from here too, I Mean, right?

That's correct. You didn't Actually talk about Yes. Uh, actually that's what Andrew's coming up to do.

Yeah. Which is Right. I'm gonna do the io.

Yes sir. He's gonna do the machine itself. Yeah.

And talk about the ROMs and all the settings from that level and stuff. Um, alright, so lemme go. Keep going real quick.

So we talked about encryption. Uh, the next one I was gonna do was basically, uh, talk about remote key management In that we support that. Again, I can't set anything at this point, but you can go into the environment, tell me where the key servers are, right?

And use this thing as a method to store those keys, right? And as we move the keys, uh, storage onto the secure enclave, we're even more secure at these kinds of certificates and keys, right? Alright.

Um, another piece, uh, that you worry about is the firmware that you upgrade the systems with. And we're, Cole and the team are talking about secure supply chain. And one of 'em is you don't want bad rogue firmware to be installed on any of the devices in the system.

And there's a ton of programmable stuff in these systems. CPDs ROMs, um, drives a nick, uh, array controllers on and on and on, all kinds of stuff. All right?

So one of the things that we talk about is that we can scan the firmware on either an on demand basis or on a regularly occurring basis to see if anything's been messed with. And the reason you have to have this is because we do scan when we turn the computer on, but what have you haven't rebooted the computer in some time a couple years. Okay.

And if somebody pushed bad stuff on there, how are you gonna know? I'll tell you how, you know, regularly scanning your firmware on regular intervals. 'cause good news, this is X 86 stuff.

It reads the firmware on post. It's already loaded into the memory. If you muck with it, it's not gonna reread it until you post.

So the idea might be don't reboot until you scan feel good about your life. You know what I mean? So, so there's a lot of strategies that that can be taken in place.

Now what's happening in the, um, technology with our partners, uh, in the industry is we're pushing the ability to scan the firmware on the devices down to the devices themselves, in that I can use a protocol, SPDM to call for the scanning and scan that bios on that remote controller card, just like I can with the processor and the other programmable devices that I, I can scan. All right? So SPDM is really important.

We're talking about it a lot to our customers. And, uh, a lot of our new, uh, nicks array controllers and so on, uh, can be folded into this such that you would scan them too. Okay.

Questions, comments, makes sense. And You can update this stuff right, right from there too. Yes.

Absolutely. There's the, this ecosystem is onboarding or loading the firmware to things, to this onboard baseboard management controller and then applying 'em into the devices. And what's kind of newish with ILO seven?

We've done this for a little while, but anyway, um, we don't, uh, in the old days we used to take a giant multi gigabyte six eight gigabyte firmware, DVD image ISO, and it had every firmware driver known to HPE going back two decades now, it wasn't that bad, but it was current machines, but still it was machines that weren't necessarily the ones you were upgrading. You didn't have all of the devices in our portfolio. So why were we giving you all of the firmware in a distribution?

My point is, in the new system, the new ecosystem, we only load up the stuff that belongs to the computer. So right there you're less vulnerable 'cause you don't load a bunch of excess stuff and you own, and the payloads are smaller. Um, and Mr.

Bradley talked in about secure gateway, which is again, an IO aggregator technique. In other words, we can put a VM in the data center, have all the iOS talk to that vm and that VM can talk to comm, right? So that you don't have all of the iOS going up to comm.

So you can do that, right? And so the secure gateway can act as this buffer or cache, if you will, and you can load the firmware and drivers pertinent to those machines on the secure gateway. And that thing can be the conduit of the iOS.

So we're cognizant of not giving you the whole giant bundle and being smarter about how we're distributing those elements. Make sense? Mm-hmm.

I'm gonna bring Andrew up and he's gonna talk about using the computer. Alright? So Chris has given you guys a great example of managing the server, um, directly connected to it, right?

We're directly connected to this IO seven. It's physical hardware that we're now manipulating. Um, so this one server, we're gonna continue down that path and try to manage it some more.

Some things that can't be done, uh, within ilo. So we had talked about a little bit before, but we do have a remote console window that we can launch and new in IO seven. It actually opens up in a new tab and ask me for a password.

So we've booted the server and got into the F nine prompt that you're able to hit to get into the system. Utilities, we're actually sitting at system help. Let me back outta there.

So yeah, we're at the just normal boot up. Hit F nine, get in here and you can do some configuration. But what are some of the things that you might wanna do on your first boot?

Right? You're gonna go in here and maybe configure some settings for ilo. You can configure some settings for, uh, your hard drives, your array controllers.

Um, and you can also configure things for the server on how you want it to perform. Uh, if you want it to be more en energy efficient or maybe you just want just raw horsepower. All those types of things can be done in these menus, right?

And again, we're just working on one server. This is the same server that Chris was showing a second ago. Um, so we can go in here and go into the system configuration, head into the RBSU.

And, uh, this has not really changed too much for, uh, you know, ILO six, previous Gen 11, gen 10 plus. Um, but we have our predefined workload profiles that we've set up already. And, um, you're able to read more about those.

We have, you know, long definitions for each what each one is actually changing. But this will help you, uh, get kind of the performance that you want without you having to know all the little dials and turns and switches and knobs that are within this system, uh, that you may not know if you need those or not. Right?

That just kind of helps a lot of people out. So for this example, I can change it to just something like virtualization max performance. And it's gonna say, Hey, are you sure you want to do this?

'cause it's gonna change a bunch of other options. It's like, yeah, sure. We, we know we, we want that.

We, we use that all the time. Are the only The those presets available or can you have your own goal configurations that are added to that list? So there is a custom configuration for sure.

Uh, down at the bottom there's a custom. And when you go to that guy, it's gonna enable you to just go in and set, you know, the one option that you needed or the 50 different options that you know, that you know you've tested and can confirm that that's what you need. Okay.

Is there a way to add that as a profile? Uh, to add the custom one as a profile? It will.

I know I'm gonna do like 30 of these or every time. Oh, you mean more than one custom? Yeah.

I don't believe there is a way Currently. Yep. Yeah.

Custom is universal through my allo infrastructure. 'cause I think she's headed on to that super important piece. If I have five settings that I need to change in the bios across a hundred machines, I don't want to Well, that you're gonna have, so the way we're showing you right now mm-hmm.

You would go to the io mm-hmm. And change those five settings manually on all of those machines. Okay.

The one, this is machine, This is managing one server. Okay. And then I'm gonna show you this and then we're gonna jump back over to comm where I'll show you how to manage what you're talking about more than one server or a hundred servers at a Time.

Yeah. There Are times when you wanna manage just one server, right? But Not as often when you're have a whole fleet of them.

Yeah. Yeah, exactly. If you only have bought one new server at the edge and you just need to set that one guy up, well, you could do it this way, but there's more than one way to do things and we've automated that to make it a little bit easier, uh, with calm.

So I just wanna show you the hard way to do it first and then we're gonna go to the easy way. You just making it hard. Well, you know what caused The manual way?

I'm sorry. Easy way. The manual way.

It's easier than the hard. Okay. He's gotta make it hurt first, then he's gotta, yeah, Sometimes you gotta boot the thing and look at the screen and type things and Well, that's what I was just thinking about it, about hugging that for, Oh, oh my goodness, you've gotten too attached already.

Did you give your server a name? So Other things that we can configure within this menu are things like, uh, an array controller. So for example, you could just go right into your array controller, uh, configure your arrays, you know, create different types of, uh, storage arrays, whether you wanna raid one, raid 10, raid zero if you're filling, uh, like you're not gonna lose any data, but yeah.

Um, this is where you would do all these things. And like you said, you would have to do this on every single server that you were setting up if you went this route, right? So, so at compute ops management, we're here at the actual server.

This is all the information and it's pulling all its information from io. Um, and one of the things that we were just showing that we could do in the RBSU is, you know, change a workload profile. You know, we have that either that custom one that we've saved or a specific one that we we know we want to use.

Uh, we can do that and add that to multiple servers. And there's a special little way that you can do that within comm. So to do that, we just go over to the manage tab, Go over to our server settings, and you'll see that there's quite a few, um, bios workload profiles already, uh, populated.

And these are the pre predefined ones that were already in, uh, the RBSU. So we can find the one that we were wanting to use and, uh, or we can just, uh, create a new setting if, if, if there isn't one that already already meets art, right? So what kind of settings could we create other than a workload profile?

You know, um, there's actually a few. So if we drop down this category, uh, you'll be able to see all the different things that you can actually create a server setting for. And this is gonna be kind of like a profile that we can apply to groups of servers.

So for example, let's look at the, uh, the internal storage of, of the server, right? We want to set some type of array that we know all the servers that we're gonna deploy under this group are gonna use the same storage configuration. So we can go over here and select, you know, server internal storage.

I'll call this tech field day storage setting. You have no idea how much of an inside joke, what you just posted is single day storage. Alright?

So once you get to here, you know, it, it allows you to do all the configuration that you would normally find, uh, to configure your arrays, right? Go in here and select whatever rate level, um, you know, whether you wanted to use the entire disc, how many drives you're gonna allocate to this array and so on, right? Once you have that saved.

So I'll just do a quick one, right? Um, save it. We have our, uh, we have our, you know, profile that we created and it's saving the setting and it has the name that we created.

Um, and now, so how do we get that on 10, 15, a hundred servers or just one server from comm? Um, and that's actually a pretty easy way to do that. So you just go back over to your manage and you can create a group.

The groups are the things that you apply the server settings to, and then you can put servers into the group and they will get all of the settings applied to them. So we can create a group tech field day group. And what type of group is it?

Sure. Oh, this is gonna be a server group. So it's asking us, so what type of settings do we want to apply to this group, right?

So first off, it asks us for a firmware server setting. So one of the settings that we could have set, uh, beforehand was, uh, a server setting or firmware. So whether we wanted to get the latest and greatest updates, if we want it to stop at a certain level, uh, depending on what servers we are in this group, whether this is, you know, a development or production group, we can allocate, you know, which firm we want to go where.

Uh, so for this one, I'm just gonna do this monthly demo. SPP setting that we have. Uh, our internal storage setting is the one that I'm gonna, that we just created.

Our tech field day storage setting, uh, bio server settings. So these are the predefined workloads that we were just looking at when we were in the RBSU. We're gonna go ahead and choose one of these, uh, virtualization Max performance was what I was wanting.

Uh, you even have the option to select a op operating system image. If you have an operating system image, uh, host somewhere that just storing a repository, storing all those images, you can add that here. Uh, also choose your ILO server settings enabled for security.

Definitely. And we don't have any external storage connected to this server, so we'll just skip that part. Click next.

Um, there are some policies that you can have it downgrade all the firmware to, you know, whatever your baseline that, that, that you set so that everybody's on an equal, uh, set of firmware. Uh, you can also have it automatically apply those firmware best baselines for you, uh, so that you're always at that level, right? So we can say, automatically apply that and there's a button that says power off server after update.

I don't want to do that. So next we have our internal storage configuration policies. We'll go through that little bit.

We can just say, you know, Hey, I wanna auto-create these volumes. It's gonna do this, uh, we can edit the volume name if we want, but for this example, there's no need to do that. And we have our bio setting.

This is gonna be the max performance. If there's anything that we need to do, we can just click down here that says automatically apply it. Next.

We can auto, auto apply the, uh, security settings. And you can automatically add servers based on tags. Uh, this is an optional part, but we're not gonna worry about doing that.

And we'll finish and create our group. So we now have a server group. It doesn't have any servers in it, right?

We got zero devices. Um, but what I wanna do is I wanna add that Gen 12 server that we were just playing with into it and show you how easy that is to do. Or we could grab a whole bunch of servers and stick them into this group.

And soon as we do that, all those automatically applied buttons that we have pressed are gonna take into effect. It's gonna flash that from where it's gonna change the bios, uh, workload profile and, uh, create the array that we asked it to. If we go back over to our servers list, find our Gen 12 server, and for fun, we're gonna throw a few other Gen 10 pluses in there with it.

We have those selection, we go over to our actions button and we can do a bunch of things here. Obviously power them on, reset them, cold boot them, uh, update all their firmware if we wanted to do that manually. Um, but we can also add them to the group that we just created.

So We go say add to group. It should give us a little list of the groups that we currently have. We're gonna go to the tech field day group, add it to the group, and one server cannot be added because it's part of another group.

What are the chances? That's okay, we'll go ahead and add it. And we just added those servers to that group.

That's all magic, right? You don't even know what's happening behind the scenes. But right now, those servers are actually getting reconfigured.

The bias workload's being changed, array's getting created, and all those different things are happening, uh, behind the scenes. It said that there, you said that there was one that was, it was already attached to, right? Yep.

Does that get configured then, or does that stay as it's No, that one will stay in its group and get the same profiles, whatever we're in there. So in order to get it out, you need to remove it from the group first, and then you can add it into another group. So you baseline every group for a specific profile.

Exactly. And for different reasons, right? For different use Cases.

Yeah, Exactly. So if we have a prod environment before we push out the latest and greatest firmware, you know, we run that on our dev environment first, our groups that we have and make sure everything's good. If there's something, you know, different that we don't like, maybe we hold off on that before we push it out to broad.

What about reversibility? Is there like a standby image in case of like, something goes bad? So there's always a few current PPS available, um, that you can pull a baseline from and you can downgrade?

Yes. Yeah. Yeah.

Like heart bleed back in the day, remember? Mm-hmm. And then I might have missed this, you might have mentioned it.

Are these settings made, uh, uh, penny reboot or is it a, uh, you have to, the systems have to be in a non booted state. Um, they're gonna need to be rebooted for those things to happen in the bios for sure. Right.

So, but the, but they'll be pending. It'll be pending until it's rebooted or you can actually force it to be rebooted during those steps, right? You can have it say that automatically apply the update part, and that's what it's gonna do as soon as you add it into Monkey.

Yeah. I per, I personally prefer to just use the cold boot feature as soon as I perform it, regardless of What's front end on the server care. It helps.

And you can pull, and you can literally cold boot that entire group, right? You can just select the group. Mm-hmm.

Click the little button and say you click A whole data center. 'cause that'd be, Uh, yeah, I don't want to see the power blip that happens when that happens. So, um, basic question.

This is all great. Terraform, Ansible, the like, can, does it, are there, you know, providers for this in any of those for infrastructures, code for item potent settings? Um, or do I have to do this?

No, you don't have to actually use the ui. Obviously we can script against this scripting. Yeah.

But what, but are there providers for the, uh, automation, you know, for infrastructures, code item potent, you know, state setting kind of tools I've bought Morpheus. Do, are you leveraging Morpheus? Yes.

Um, I'm not sure I have the right answer for you. Tell The declarative stuff, right? You know, I make my thing look like this as you know, 'cause obviously very, it's very unlikely as much as you guys would love it, that the whole place is gonna be HPE, right?

So, or So with the ops ramp part is kind of, I think is where you're trying to go, is like, are we able to do all this with other, Are you, I will this participate in, if I've got a, a customer that has multiple different, then multiple, you know, maybe he's got, you know, some storage from another vendor or he is got whatever net Cisco networking and they, and they decide they're gonna use Terraform or Ansible or these other things to, that's how they're gonna run their environment. But this is all to manage the HPE fleet would be great. Is there a provider from Terraform or Ansible to participate in that?

Yeah. Or even if I just have infrastructure to code, it comes up to, uh, how do I do, uh, profile lifecycle management if I want to do, get, um, repositories, et cetera. Yeah.

To just manage my, you're giving me the ability. You get the script Script. Yeah.

The even manage the script. So how do I, how do I manage the state of my right, Right. Instead of it's my code For This instead of do this, make it like this, right?

Yes, exactly. Yeah. So there are a couple of ways that we, we, uh, attack that particular question or that opportunity for automation.

Uh, the first one is obviously with tools built into GreenLake such as Morpheus or Ops ramp, however, a, um, kind of a community driven approach of creating integration points. Um, I think of, uh, something I mentioned a bit earlier is the brand new power shell, um, command lit library. It interfaces directly into compute ops management.

That ecosystem continues to build out. Um, so there are a number of entry points, but what's core to all of that is basically a a an API, uh, with published, you know, a published API that can be leveraged by those integration points, Right? So I mean, it, it comes down to just has, are there community based?

Um, are there, are there publicly available or perhaps mm-hmm. Maybe third party adapters? Or are you developing because it's difference between, you know, imperative versus declarative, right?

I mean, this is all imperative and imperative. You know, you're doing something when there's a human in the loop like that, right? There's no test.

The testing is different. There's a lot more errors. You can click on the wrong thing.

Maybe your browser doesn't, you know, there's some error, right? When it's declarative, it's make it like this and it's just, you know, it doesn't stop till it gets like that or it, you know, returns. It's just a different world.

And a lot of people, a lot of customers especially of scale have moved to that, um, whether Terraform or, or Ansible or the like, right? So, um, I would ex I I, I'm kind of surprised that wouldn't, I wouldn't be surprised if some third party HPE consumer has written an open source, you know, adapter using your APIs, right? If you've got APIs, it's doable, right?

So I'm surprised that if it's not available, I was just thinking, is it available from a third party? Are you guys working on it? Is there, talk about it.

Have these conversations all the time. Okay. And that's, that is really the, the catalyst that brings out the, the PowerShell commandlets and uh, other tools.

I think the ecosystem's growing from different language bindings and or tool sets. So it would not surprise me if terraforms on the, uh, on the horizon Or even before following up that is there like an et CD configuration drift management. I mean, 'cause once you get to these profiles, what does he do to skew right?

You know, declared it was one thing, but Configuration skew is Andrew's middle name. I don't know. Your mom was really strange.

No. So yeah, definitely calm. That's what, that's really what calm is for, right?

It's gonna be able to tell you all those things. Like if you have certain things that are out of the baseline, if things are out of place and you want to get those all back to where they need to be, you can definitely have that, uh, kind of automated for you, right? You don't need to take that action for you to manually go in there and update the one guy that's out of date or you know, the other one that has this, this weird delta, right?

It's gonna take care of that for you. Uh, by managing with those groups Over the years, we have been working on, uh, data center management for a long time. And we believe that we've really got the onsite management solution thing, uh, at a very mature and we think a very good product.

It's called OneView, the product, uh, ships as the primary control plane or control software, if you will, for a solution product portfolio of ours called Synergy. And we get asked all the time, gee, all this calm stuff seems so new and fantastic, I guess you'll be turning off that synergy stuff, right? I think OneView stuff, right?

And the answer is no, no, we just announced gen 12 for synergy. And there's other things on the roadmap. Uh, and so we're committed to that platform and that platform is managed by OneView.

Uh, many of our customers think the cloud management strategy is awesome. They might be able to move to it over time, but they use a very robust onsite in-house sometimes, but onsite management strategy. And they still expect, uh, us to support OneView for a while.

'cause that's what they're gonna use. And they're not gonna go to com right away. Okay?

So, so we're still gonna have one view in the portfolio. You all remember one view, right? Remember, I thought it was the thing until It is the the thing, Keith, come on.

It's still the thing. And in fact, we can find it here in com. Um, I can, uh, go to servers and support, you know, look for 'em.

Here's my synergies right here. And I think I can click through to this. Uh, I'm gonna try this one right here.

All right. Uh, and there's my appliance and I should be able to launch this thing. All right?

So right away y'all noticed that the IO seven looks a lot like GreenLake and comm and that does not, 'cause this is the previous interface that we did and we've moved to a more modern one for comm and for that development environment, right? So this is the environment that you still, you know, you have your servers and you create, um, your server profile, um, configurations, standard settings. It has all of the robust settings that you saw just a minute ago.

'cause again, it mines ILO for everything you can do to the computer. And then you take all that stuff, put 'em into what's called a template, and then apply the template to the servers. They configure that as a profile and that's how you replicate settings out.

So I'm not as up to date as my peers are on this. So I'm, if I'm not at up to date, I'm sure some of the folks watching. Is it, when did one view get kind of subsumed by con?

I've never even heard of it, quite Frankly. So it's not sub What, because I live in this, I live in one view, right? Right.

And that's because For my, for Myra storage and all of that between GreenLake and OneView for my deep data insights, I come here, Right? And the strategy on OneView was you installed it yourself inside the data center mm-hmm. On an ESX cluster.

And then that thing probably sat next to other management apps in the data center. So maybe vCenter's on that thing, right? And other stuff.

And every time I rev this, which was a couple, three times a year, I'd let you know that you ought to come in here and add this functionality to pick up new service support or whatever the heck it is that we're doing in here. And you'd say, yeah, that sounds awesome. HPEI should really get to that.

Mm-hmm. And within a year or 18 months, you might roll out that change to this. So you've taken it on yourself to keep the thing up and running to keep it current, to maintain that clustered infrastructure to run it.

We thank you for that. It scales to 2,500 devices. We've recently made it more than that, but that was kind of our initial scaling measure.

So if you're in a large data center, you started ending up with what we called one view islands, one view for this. And by the way, one view for Proliance DLS and and so on, uh, MLS and so on, is not the same as the one for synergy. So Synergy one view doesn't run the Proliance and ProLiant one view doesn't run the synergy.

You have to add two for each one. So it's two views. It was, yeah, two, two views.

Uh, let's not be open with those views. Just kidding. Boy.

See if anybody's awake still in the afternoon. Uh, and nobody was. Oh, we got it.

We got, alright. So anyway, uh, yeah. So, so that's the model, the scalability, how fast we can update it, right?

Taking away your need to provide that clustered server environment to support it. Can we make the workload yourselves of managing the environment less for the customer? And so it's not a replacement because you need both.

But if you're a government data center, you're not gonna go up to comm. We hear about that all the time. No, we're not gonna put our machines on the internet, blah, blah, blah.

They gotta be, you know what I mean? So you have to have an onsite management solution. And many things still run that way with some kind of vm.

So I, I mentioned, um, the calm gateway, secure gateway that we just came up with, that's a little VM that would run on your ESX host. That's to keep all the ISOs from having to be plumbed out to the internet. You can talk, they can talk to the gateway and the gateway can go out to the internet, stuff like that, right?

So there's gonna be an onsite strategy, there's gonna be a cloud strategy and we think, uh, many customers are gonna do both, right? Maybe the data center's got some one view they're doing comma and the remote sites, and then they're growing to one or the other solution depending on, uh, what their needs are. So A little aside, the gateway, you can deploy the gateway in the field as well, I think.

Yeah. Oh yeah. You could absolutely Play the gateway anywhere that's Yeah.

Anywhere you got an ES x the value of that. Yeah. Have an es it's a vm.

Yeah. Can you deploy the gateway and install one view on the gateway so that no Keith can The, the ESX host that the gateway has got a little VM running on right next to it could be a one view right next to it could be V center. Yeah.

It's just a little, you know what I mean? It's a single function little vm, right? So anyway, the point is that we still, this world really does count, as you can imagine.

Um, yeah. It's, you know what I mean, it's not necessarily, um, the newest thing. Um, you know what I mean with the buzz and all that, but boys, I like to call it bread and butter.

'cause you know what I mean? That's what you run your data centers on mostly right now anyway. Alright, well, So the customers are trying to get out of the data center.

You know, I don't want to add the service. I think this can be leveraged, right? I mean, oh yeah, you can, MSPs can use that, what have you.

Oh yes. And Comm Comm has got an inherent, um, managed service provider construct where you can sign up, um, and, and subdivide the Yeah. The workspaces for, you know what I mean?

And MSP can take the top one and subdivide the lower ones. Yeah. That'd be your different, So vendors Instead rack, right?

Mm-hmm. Mm-hmm. Anyway, um, I think that's the content I had.

We do have 11 minutes left that we'd like to leave that open for Mr. Schaeffer to answer any questions. Scott.

Sorry dude. There, there's a bus that was coming down the road. Anyway, please.

So let's talk a little bit about Lifecycle, because at the end of the day, this is where the value is as customers are trying to figure out. And especially as we bring more of these beast of, uh, machines in with, uh, um, with GPUs, how are you helping us figure out how to manage our life cycle? Because that's the challenge.

I, I can only depreciate, or I can't even depreciate GPU hardware anymore because NVIDIA's, uh, life cycling this stuff out every 18 months. And I am, the, the, the speed of refresh is just something I can't keep up with. You don't have to keep up with the Joneses.

No, but that's, That's a good point. So, you know, obviously there's, um, higher level, um, IT service management tools that we can plug into. Uh, we do have the REST interfaces makes this quite programmable.

So depending on your lifecycle management strategy and tools, we sell 'EM and ops ramp and other things, uh, and we provide our ability to plug into them. A good example is, uh, plugging in a vCenter capability. And if you wanna run out a V center, we can, you can drive our environment quite a bit.

Calm, uh, the Islas and so on. Um, so, uh, I would say that, um, for a small consultative fee, HPE services would love to come out, sit down with you and organize and talk about rolling out a lifecycle management strategy, uh, that we would help you with for the next decade, because that's how we keep in Business. I was about to say the next decade.

Okay, so that's, I'm throwing out my GPU in 18 months. Yeah. Really.

But, but we seriously have consultative sailing selling motion and that kind of aspect in our services organization that completely focused on lifecycle management. Uh, and we publish a ton of, uh, supporting, um, white papers, content and so on. You want me to say something on that?

So Keith, were you specifically talking about GPUs though, or are you speak, I thought maybe you were speaking more generally. So Are you speaking the, whenever I was in charge of a, a hardware set, the stuff moved faster. Even pre GPU it moved faster than, than I would like because, you know, my people can only get up to speed so quickly.

My people can only pull out servers so quickly. And my app team was never ready to deploy a new app on any platform, let alone, uh, be of any assistance period. So any, any optimization I can get in the whole lifecycle pipeline, whether it's, uh, being able to take settings from my, uh, believe it or not, there's still a lot of IO Gen five out there.

Oh yeah. Taking that stuff, translating it into IO seven and helping me with the engineering effort to get to a, uh, just operational functionality of migration. It is helpful.

Yeah. And I think, I think we're saying, right, one of the benefits of ops management com is really providing that management experience across the, the generations. So you don't have to throw out your gen tens if you don't want to.

It will still manage them. Oh, I Don't, let's have Them. Yeah.

And, uh, uh, now at some point you should though because they're consumer more power than the value you're getting from the, uh, workload That's somebody else's, that's somebody's budget. Um, That's still over time. Yes, but right.

You know. Um, but, but yeah, that's, that's how we're thinking about the problem is that customers will have a mix of generations of systems and we wanna provide as common experience as we can. Of course some capability is only available from the, the latest generation, but that's okay.

Yeah. So it would be interesting to see like, uh, and even from a capability perspective, like you mentioned, power, power is, I don't necessarily expect to get power out of ilo. Maybe I do, I don't know.

But one of the things that helps justify upgrades is being able to say brand green. Oh, this group of Gen 10 servers compared to if I had a group of 10 Gen 11 servers, here's the changes I can make at the IO level. They'll make me more efficient and my environment more efficient.

So I can justify going from gen 10 to gen 11, whether it's, uh, cost of operating the systems from a power perspective or even man hours that I've spent on firmware upgrades and this ability to, to do things that I can do in Gen 11 that I can't do in gen 10 that I can now report on in my environment to take to my man who's typically me to, uh, to justify That. Yeah. I think that's, that's good feedback.

Um, and We were Gonna show the sustainability on green line. Yeah. And uh, uh, but it's interesting you say you weren't expecting to get power, uh, out of ILO and you should.

Yeah. Right. And that's one of the things that, again, ILOs individually and comm and aggregate, is pulling together so that you have that view of how much energy you're consuming.

Now we're approaching this through the lens of sustainability. Now appreciate that some people, like, especially in North America, that hasn't been a major concern, but in some other parts of the world, Europe, it's a significant concern. And so we've been working on providing the visibility into what your energy consumption is, uh, so that you can collect that data, turn that into your carbon footprint if that's what you need to do.

We actually make it really easy to turn it into the carbon footprint, right. In the tool so that you can, uh, include that in any of your in, if a customer has to produce a sustainability report, for example, and we have that data right here. So for example, you can have your carbon emissions.

So we calculated that based on the amount of your energy consumption, your actuals, and we turn that into this sustainability report. Can I, can we go talk a little bit about the other part of the lifecycle management is not these types of devices, which are the edge devices, right. And fleet management, um, in the edge.

So a lot of, a lot of what you do in traditional data center, you have a very, a single or a few locations of a very large fleet of machines. And the edge is the exact opposite. You have one or two, many, many, many locations.

You could have a thousand locations, each of which has one or two, One node, right? Right. Absolutely.

So it's a different way of looking at a different way of thinking about and managing it. And one of the challenges in that environment is bootstrapping, how do I get my systems out to those many locations? Do you have or have you thought about a way to do that where I don't have to buy a bunch of machines, have 'em shipped to a central location, configure 'em, and then ship 'em out?

Yeah. Where in fact, I can ship 'em. Yeah.

I can tell you build ship to 14 locations, a hundred locations. Yeah, absolutely. In fact, that's one of the things we were kind of showing here, is that you can deploy the server, connect it, it onboards into comm, but deploy the server at the end location at the edge, right?

It onboards to comm and then you can say, okay, I need you to deploy this image to it and get it up and running. In fact, we actually worked with a customer to build a solution where that's all automated. So as the servers come online, it automatically gets their OS deployed and their application stacks and there's no manual intervention required.

It's just go in cable it up, powered on it, connects to their network and off they go. Yeah. Andrew, shoot, it was a setting, one of the server settings was the boot, um, os deployment.

Well, yeah. And I'm even thinking, so, so a lot of the situation is I'm, I'm gonna, I'm gonna say deploy to a hundred retail locations, right? I'm, I'm now redoing my thing.

I'm gonna deploy to a hundred different retail locations. Now how do I go and say I, you know, maybe I, I don't have A-D-H-C-P server at that location, right? I have nothing other than the previous box there had a fixed IPI address, right?

And I wanna ship a box to that location that gets that I that, right? So I don't have, all I have is a guy who's gonna put it in, right? Come put it in, plug it in, and that's it.

Right? Yeah. Walk away, right.

And The lobby secretary, Right? So, So he takes, so At some point there, this is what we worked with this other customer, right? At some point there's a, you have to start making some security trade-offs, right?

You have to decide what you're comfortable with, right? If you're comfortable with the server automatically doing some things that, that would bypass security, then you can do that. I don't may not advocate for that, but I appreciate the challenge because as you just said, I don't want to have anybody knowledgeable out at the other end.

Well, it's not so much that I don't wanna have any knowledge, anybody knowledgeable. Well, it's, I don't or you don't. Right.

I don't have those people. I'm particularly in a, in a retail location, that box is going to be shipped to the store manager. That box lives in the store manager's office.

And what you say is take the old box, unplug it, put the no box in, put the old box into the, into the shipping box. This is the right, get it Back. This is the set of trade offs.

You have to decide because Yeah, absolutely. And what we've been thinking, what we think is that for some customers, for which security is paramount. Yeah.

Okay. Then what you should do is you should ship them to a central location where one person can configure the security trusted. Mm-hmm.

Set it all up and then you deploy it to the end where nobody knows anything, they just plug it in and go. But if you, if security is not paramount for you mm-hmm. And you wanna be able to just ship a box directly from us, it comes right out.

Our factory goes right to the edge location where nobody knows anything but you wanna plug it in. Mm-hmm. We can do that too.

Okay. Because we can set the IP in the factory. Right.

So, so that, that was, that was sort of where I was leading to is Right. That's a really useful fci uh, service where I can basically say, Hey, I need a box hp, can you just ship it there because for you to ship it to me, me to do something with it and to ship it off to the location, that's time, money, effort. People that I don't need to do, if I can just say, Hey, We can, this IP address, we can ship it out, we can pre-configure it before it goes to that end location and then it connects up.

Yep. And you're ready to Go Basic Composer as a service. Yeah.

What we're doing here, kinda, you're gonna, you're gonna boot up and you're gonna Mac address, advertise across an IP link and then you're gonna be able to download Composer component functionality. Well, Yeah. You're gonna be able to, able to connect it up.

Yeah. Right. Right.

That's, I mean that's, that's really a key for the Edge Hyper and, and I choose retail 'cause it's odd people understand it, but it's all the different types of things where you have, like I said, it's the inverse relationship where you have a huge number of locations with a very small number of devices And that, and that's one of the big benefits of comm. Right. Our whole point of that is that we, you know, before if you had one view and that was all we had, you somehow have to put a one view at every single retail location.

Right. Right. It's just ridiculous.

Right. It makes no sense. So instead it's no, I want one cloud-based management solution where all of my individual nodes connect back to it and I don't have to manage the infrastructure, I don't have to do anything.

Right. Right. It automatically manages my thousand remote locations.

Um, exactly as if they were in a data center. In fact, I defy you to know the difference, right. When you're using the tool, if it's local or remote, other than the map tells you.

Right. Right. Yeah.

So very important. Yeah.