AI Ops to Edge Management - Simplification of IT in the Data Center

James Hilliard:

Welcome, everyone, to our Ask the Experts event, AIOps to Edge Management: Simplification of IT in the Data Center. This conversation is being brought to you all by Connection in partnership with Intel^® and Lenovo. My name is James Hilliard. I'm glad to be hosting this conversation. Three guests join us today. Gretchen Stewart is here, the principal engineer and chief data scientist at Intel. Gretchen, thanks for your time.

Gretchen Stewart:

Delighted to be here.

James Hilliard:

Also with us is Jay Bryant. Jay is a principal architect in the ISG software and solutions team there at Lenovo. And glad to have you here, Jay.

Jay Bryant:

Thank you. Glad to be here.

James Hilliard:

And last we have Chi Chung, the director of solution architects at Connection. Chi, good to work with you again, my friend.

Chi Chung:

Yeah, thanks for having me.

James Hilliard:

Absolutely. So what we're going to do, folks, a couple of questions up front just to chew on this idea of how complicated the data center has become, and then we're going to really open up the mailbag. So we've been getting a lot of questions from all of you, a couple pages worth. I hope we can get to most of these questions in the time allotted. So we'll go through these, we'll ask the experts, at the end we'll give you some contact information so that you can reach out to the Intel Lenovo and Connection teams to learn more as you need.

So that's what we're going to be doing. I want to start with a bit of a controversial question, if I may.

Jay Bryant:

Uh-oh.

James Hilliard:

Here's the question, I even wrote it down, controversy, I want to make sure you get things right. Is the data center complicated today, yes or no?

Gretchen Stewart:

Yes.

James Hilliard:

Jay, is it complicated?

Jay Bryant:

Absolutely.

James Hilliard:

Chi, is it complicated?

Jay Bryant:

Definitely.

James Hilliard:

Now, let's chew on it for a while. Gretchen, how did it get this way? And it's not the first time we've seen some technology start off really, really useful and simple and then really sprawl.

Gretchen Stewart:

Yeah, I think the first thing is people need a strategy, and they really don't have one because today you can think about going to Equinix to have bare metal servers. You could use the systems that you already have. People are just adding without thinking about it, and you really do need a plan, especially today.

So when people are really thinking about the baseline or consolidation, they could be looking and talking to Lenovo and say, "Hey, we already have 10 or 15 servers. Could we potentially put those into two or three?" And then when you need to look at the software, you've got Connection where they have solution architects and people who can really look at how much Microsoft do we have? Do we have too much VMware? Oh, by the way, there are different softwares that came out at different times that overlap each other. And so it really makes sense to literally step back and go to the basics and start with looking at what you have in your data center and does it make sense to grow your own data center, or again, think about Equinix and you can get bare metal servers and just basically take your gold image, move it into their server farms and be able to manage your entire data center.

James Hilliard:

Sometimes I feel, Jay, that when we do go through these ebbs and flows in technology, sometimes when we get complex it is because the speed of change has just been so great that we don't give ourselves enough time to stop and maybe think of the whole picture. We get a little narrow focused. That's my observation, but how did we get so complicated again this time in the data center?

Jay Bryant:

I think it's because customers are thinking about the things that they want to do, what are the applications they want to run? They want to focus on the problems that they're solving, not on the infrastructure that exists to try to support that. And so what my group has been doing with Lenovo is really looking at how do we make people's lives easier so that they can focus on the cool stuff, the AI, the problem solving and not worry about their infrastructure and giving them those insights so that the infrastructure is there and it does what it needs to, but they can do the other jobs they want to do, not managing the hardware.

James Hilliard:

Chi, what do you see as some of the things that have led to today's complexity in the data center?

Chi Chung:

Yeah, absolutely. I think there's a lot of silos in IT organizations.

Gretchen Stewart:

Still.

Chi Chung:

You have virtualization teams. You've got networking teams. And within each of these teams, they've got their own list of initiatives that they're being asked to deliver. And the reality is they're not spending the time sitting down and saying, "Hey, I'm doing this. What are you doing," and sometimes the plane doesn't always land at the same time. And as a result of that, we have all this sprawl. No one's really looking at it from a holistic point of view, which is really where we come in right, along with Intel Lenovo, where we really take a look at the big picture and say, "How can we improve your experience," and ultimately I think that's the challenge that a lot of IT organizations are facing right now.

James Hilliard:

I want to talk about the customer base out there, the teams that rely on your teams for this, ultimately we're talking about here simplification. What is the biggest pain point? When you engage with some teams that are saying, "Our data center, we have been sprawling. We've got these silos still," is that the biggest pain point? Is there something else that they point to, and I want to hear what you hear from your customers as well?

Chi Chung:

No, it's a great question. I mean, from my lens, what my team sees when we talk to customers every day is there's a lack of skills and challenge with finding the right resources. The IT industry is changing on a very rapid pace and most of our customers just don't have the bandwidth to keep up with their day job while also learning about all the new technologies, new modern ways to do things, and that's where we come in. We really want to sit down with them and show them what the art of the possible is because they may be familiar with the way they've been doing things 5, 6, 7 years ago, but things have changed quite a bit. And this is where if IT organizations are open to change and receptive, then this is an opportunity for them to make some adoption of new technologies.

James Hilliard:

Gretchen, what are the pain points that you're hearing from the Intel customer base, the folks that you get engaged with, where are they saying it hurts and they're looking for that change?

Gretchen Stewart:

I think for a lot of the customers that I'm talking to, they are wanting to solve those business problems, but they also have the bright shiny object stuff that's going on. And so honestly, I'm a math nerd and then I went back to school and focused on data science, so AI is the answer, what's the question is what's going on today. And the truth is, to Chi's point, that you ultimately can be just using remote process automation to help them do their job today. That's something that's been around a long time.

There's a lot of predictive analytics. There's a lot of these are the kind of conversations that I have in terms of, "Hey, we have some challenges within our network." Well, we can do anomaly detection. That's just standard AI that's been around a long time. It's not generative AI. We can help you with leveraging that, but it's really using that technology to not create a forklift change, but really to try and take what they have, again, whatever that baseline is and leverage the technologies.

I think the one thing I always say to people is that, and I'm a crazy baseball fan, that's why I love math, is that you don't need 10 pitchers. You need everybody playing their position.

James Hilliard:

True.

Gretchen Stewart:

And everything that we do in the data center, everything that's related to AI is a team sport. And to your point about breaking the silos down, people really need to understand that it is a team sport. You cannot do this by yourself.

James Hilliard:

Silo is one of the biggest pains that you hear?

Jay Bryant:

I'd agree, and I like what Chi said about that because very often in the larger data centers I'll talk to customers and it's like, "Okay, we need this network configuration to help support this." And they're like, "Oh, I need to go talk to the network team. I don't know that." That's a real challenge.

But the other thing that we're seeing with Lenovo is it's not just the big data centers that we need to help. It's the small to medium businesses where IT is becoming absolutely essential to them, but they don't have data center admins. It becomes programmer A, B's job to take care of their own hardware, and we need to help them with tooling, visibility into it, integrated information and education so that those smaller companies can also leverage the power that's now available in IT.

James Hilliard:

All right. Here's what I want to do now. We've set the stage, and yes, we know what you know and what you experience every day. The data center is complex, and so we want to talk about the simplification now, which is what a lot of your questions started talking about. So that's where I want to start with the questions from our audience out there.

First one was from Brian, and Brian's saying here, "What strategies are most critical to simplifying IT complexity," because we know there are a lot of strategies. One of them, Gretchen, I want to start with you. You were like, don't just focus on the shiny object. That might be a strategy, but your thought, the most important strategy because there are a couple, but what do you put at the top of your list on some strategies, we'll talk about tools and some things later, strategies IT teams can take to start a simplification journey?

Gretchen Stewart:

Yeah, I think there's probably three things. The first being cost, think about what your cost is today and potentially what your budget will be in the future. I also think it's really important to think about the user experience and what are your SLAs that you really want to meet for your users. And then the third is security, how are you going to do all of that in a secure environment, especially depending on what your industry is?

James Hilliard:

Chi's smiling over there because security is always one of the top things. So that's got to be your top. You got to focus on that. What else?

Chi Chung:

Yeah, absolutely. I think the approach is taking a crawl, walk, run approach because one, depending on where in the journey you are, it's going to take time. And the reality is you want to rely on, you find partners that are familiar with this space, folks who have done it before who can help provide those expertise, that's gone through the trial and error, if you will. Because if you do it on your own, your speed to market is going to be a lot longer than if you work with a great partner or a great OEM partner like Lenovo or Intel. So my suggestion is don't go at it alone. Make sure, give yourself plenty of time to be successful with it.

James Hilliard:

On the strategy of crawl, walk, run, there are some IT folks that are getting the pressure to just run. Any advice on how they might be able to slow that down when talking back up the chain of command to their executives? But I know you want to run, I know we need to get there, but we have to stop and slow.

Chi Chung:

So this goes back to what Gretchen was talking about the importance of a strategy. You need to get your leadership buy-in. You have to get their support because if they don't understand what it's going to take to get there and what the vision is, you're constantly going to be put in a reactive position. And really if you can get leadership to buy-in, you can get the right stakeholders to support you, then they're going to help you slow the train down and give you the best chance to succeed.

James Hilliard:

Okay, fair enough.

Jay Bryant:

One of the things I like there that I think we can communicate to leadership by having good processes and strategies to explain why we need this. If you don't have a good process for bringing your hardware on board, setting it up so you're getting the data from it, having consistent methods for applying across those systems, you're going to just lead into a nightmare and trouble of management in the future. And so bringing together not only process, but then using that to keep your hardware running, your systems running the way you want to and being able to explain the importance of managing and using that hardware.

Chi Chung:

Efficiency plays a big part of this, right?

Jay Bryant:

Yes.

Chi Chung:

We're really talking about efficiency. Because if you're not efficient, and especially to the stakeholders, to leaders, if you're not efficient, you're wasting money, and that's something that I think always lands with leaders. They understand, hey, if we rush this, we're going to end up spending a lot more resources and money to get this off the ground.

James Hilliard:

And that's the efficiency that the business side understands is on the monetary side. There's a different efficiency though on the technical side.

Jay Bryant:

Correct.

James Hilliard:

Right?

Jay Bryant:

Correct, and I mean, I think there's an impression that, and there are a lot of smart people in the IT industry, and we're solving some crazy problems, but there's always this push to do more faster and you end up if you're pushed that way, you end up making mistakes, you have to go back and do it again, and then ultimately you're less efficient, you're wasting money.

Gretchen Stewart:

Yeah, and I think having people like Chi and myself and Jay is that we have experience and that you really, it's great to have somebody, and I don't want to be ageist, but it's really good to have somebody out of college, but also working with somebody who's got the experience. Because I think about as an example at Intel, we have different IT rules depending on if it's a fabrication plant or it's folks like myself in the industry just working and using the applications, and that's true even for small businesses. You've got somebody who's trying to be more competitive on their shop floor. There's a piece of that, but then there's also their office workers or their salespeople who are using different tools and really having somebody who understands that it's not we can do, as I call it the spray everywhere and pray it sticks.

Jay Bryant:

Spray everywhere.

Gretchen Stewart:

It's really let's make sure that we understand all the components and then really build a system of systems because that's really what the data center is.

James Hilliard:

We've touched on some of the strategy. Shirley, question, is what about then the tools out there? Because you can have a good strategy, but obviously a strategy, the next level is, okay, there are tools. So her question, "What are the most effective automation tools for managing large-scale data center operations?" Again, a little bit different too because we talked about we have some smaller companies out there, but let's start at the large level. Maybe Jay, start with that. What are some of the tools out there that people might want to be exploring, once they have a strategy, once they have a plan, once they have executive buy-in?

Jay Bryant:

So I mean, we could spend the rest of the day sitting here talking about what tools are available. There's Ansible, there's Puppet, there's Chef, and a lot of the answer to what is the right tool depends on what people you have to develop those tools. Going back to your earlier comment about you have to have the people to implement it and you need help with that, you can take these tools and do it yourself. And one of my coworkers had a restaurant analogy, are you going to be a McDonald's or are you going to be a Michelin-rated restaurant? If you do it yourself, McDonald's does it good, but if you go and you get help from partners like Intel, Lenovo, Connection, we can help you build a Michelin star restaurant.

And so then it's a matter of what do you need us to help you do and what tools fit that need? And at Ansible you can do it, and it takes, it's not the fastest. We've talked to customers that went and used Puppet or Chef instead and we're like, "Well, why'd you choose that?" It was faster. They felt that Ansible took too long. So there's also a process of investigation and discovery that has to happen there. But when you work with people, like you said, who've had the experience, we can help guide you in a direction that will maybe save some of that time and investment and research.

James Hilliard:

So what Shirley was asking, what are some tools that you would suggest folks evaluate, think about?

Gretchen Stewart:

I think we put together Ansible playbooks all the time to hand out to customers to be able to use, again to make their life easier. But some of it is that being an engineer, working for an engineering company, we build some of the most sophisticated hardware that has intrinsics built into the chips themselves or built into other components, whether it be networking or silicon photonics that we build it and we truly believe they'll come.

And so part of why we've tried to start to do things with some of the tools that you said, but even some of the applications like SaaS or VMware or others, is expose them to all of those intrinsics so they're going to just run better on our platforms so that it really makes life easier for them. So if they're trying to find a way to ensure that they understand the security and that they want to have a trusted execution every time they reboot their system, well, we have a trusted execution components that are built into the hardware. And if you're using these tools, then it's already there and it can absolutely that it's a hundred percent guaranteed. If you use these other tools, it doesn't. So some of it is us also having to work with the tool vendors.

James Hilliard:

Let me, since we're still on tools, I'll mix another question here, this one mixing in the idea of any other open-source platforms to be looking at to automate the data centers and operations and all that. So the open-source front, is that any thoughts, tips, ideas there?

Jay Bryant:

One of the tools that we've used internally in Lenovo and built some solutions around is called NetBox. It's a good open-source tool for doing automatic IP address management, onboarding your systems, getting the information about all of your systems in there, being able to search it, that kind of thing. So that would be one thing. We've talked about Ansible. Those are the big ones that I have been working with most recently. Chi, did you have additional thoughts on that one, or Gretchen?

Gretchen Stewart:

There are some things also built into some of the operating systems. So if you're using RHEL or you're a Windows environment, there are things already built into those. But at the same time, you want to make sure that those connect to whatever dashboard product or other performance tools that you might want to use. So again, that goes back to the having conversations with Lenovo and Connection to really figure out what integrates with each other before you start buying all of them, which this has 10% of what you need. This has 20%, but this one already also has 5% of what you... It's planning.

James Hilliard:

It's adding to the complexity is what it's doing.

Gretchen Stewart:

Exactly.

Jay Bryant:

And the interoperability is really key, and I think that is also a call to action for all of us here. There have been many attempts in the industry, whether it be Redfish, OpenBMC, et cetera, to reach standardization, but it always seems like somebody just works around it a little bit so it's not totally standard. And I would like to see more of a commitment from the industry of working on making things more interoperable, better, more consistent standards.

Chi Chung:

When we work with clients, one of the first things that we do talk about is do we go the open-source route, is it the right path? Because again going back to some of the things we already talked about, it's part of the requirement gathering process where ultimately you may decide that, hey, for us to get to where we need to go, we may need to use a solution from a well-known OEM that's going to get you there quicker and you're going to get that support. Larger companies tend to look at open source, and then for small to medium businesses that quite frankly doesn't have the resources, doesn't have the skills, that's where they lean on partners like us and some of our OEM partners to really help bring those expertise and help them get to where they need to go.

James Hilliard:

AIOps, obviously in the title of this Ask the Experts, there are some folks where especially I'm going to suggest smaller organizations that might not be there yet, Chi. So can you do a quick, and then I want to get into a couple of questions in here about that, including Evan and Casey that wrote into us, but a quick primer, the one-minute version of the promise of AIOps for teams out there.

Chi Chung:

Yeah, simplifying operations, better visibility, better control, more efficiency, and the ability to flip the script. Imagine more proactive IT organizations versus reactive, being able to go to the business stakeholders now and say, "This is what we're using. This is what we anticipate we're going to need, so this is what we're going to budget for." Whereas in the past it's been very much, hey, throw a dart or whack-a-mole, right?

James Hilliard:

Same as

Chi Chung:

We saved back whack-a-mole a lot.

James Hilliard:

That was good.

Chi Chung:

How was that? Was that one minute?

James Hilliard:

That was a minute. You did well on that. All right, so now that we have defined AIOps, let's talk about implementing it, and sometimes teams might be ready today to implement. There may be teams that aren't ready yet. What are some things to think about in terms of implementing that? That's one of the questions that came in.

Chi Chung:

I talked about silos earlier. You've got to break down those silos. You've got to communicate amongst the team. When you start talking about implementing AIOps, a lot of it is around metrics, what are the SLAs, really having an understanding of that. And ultimately making sure that it aligns back to the business stakeholders too because sometimes IT has different standards than what the business thinks. So I would say definitely get everyone together, really make sure everyone understands the mission and break down those silos and communicate.

James Hilliard:

What do you say? Follow up on that.

Gretchen Stewart:

I agree. Well, I think the other thing is that again, people think that generative AI and things like that can really make and save the day, but to your absolute point, once you've broken down the silos, it's figuring out what is that data that you've already been collecting and is that data data that's going to be helpful, or in some cases it's probably stuff they should have thrown away a long time ago because they aren't collecting the right information.

So part of it is really also sitting with the business stakeholders to understand, okay, there's the operations, but there's the, hey, I sit around and wait for five minutes to open up this particular application, and when we start adding that with all the people that are waiting, you know what I mean? So it's this combination of are we collecting the right data and do we even really know what data we should be collecting? And then once you've done that, automating it is really pretty easy.

Chi Chung:

Well, we see a lot of clients where they implement solutions and then they spent the next month or so trying to figure out where the threshold is because alarms are going off left and right because they didn't go through that process of saying, "Hey, where do we need to be?" Because if you don't, everything goes off and now you're just spending all your time triaging, which ends up actually going against what you're trying to do in the first place.

James Hilliard:

And goes back to something you said earlier that you don't want to just have all of these put it out there and then now, oh, wait a second. We were running before we were crawling or walking there. And implementing AIOps, it's something that the question in here, what are some of those considerations? So what's something top of mind for you on that?

Jay Bryant:

From an AIOps standpoint, first, and I sound like a broken record, but you need to have the data coming from your systems. You need to have a plan to get that data saved. And I know that Lenovo has the ability to get that information there, but you need to have the data to be able to get the insights on it.

The other thing that has caught me off guard, as I've started trying to use AI in my daily life, is you also have to have the resources. I mean, if you're going to train a large model, you need to plan to have the resources available to do that model training and then be able to do the inference against it.

One of the thoughts that really struck me as I was planning for this and thinking about talking to you guys is this is also a call to action for Lenovo, for Intel, for Connection, for our partners, that with AIOps, the AI has to know what a problem is, and we don't want the customer to have to have a massive failure in order for the AI to learn, oops, that was a problem. This system was about to keel over.

And so I think it's an opportunity for us as partners, as providers to our customers to start looking at we've seen that happen in our testing, provide that to our customers so that they don't have to have that pain. I think that an area that's going to be very interesting going forward in the AIOps area, and I'm excited to see what we can do there.

James Hilliard:

That knowledge share out there in the community. Casey had a question, "What security measures should be in place to handle AI-driven systems?" Security's a big deal. Talk to me.

Gretchen Stewart:

Security's a huge deal, but it starts at the beginning.

James Hilliard:

It's not the add-on at the end.

Gretchen Stewart:

It's not the... Yeah.

James Hilliard:

Which it used to be way back when we all started.

Gretchen Stewart:

That's why I say to people you need to build it in and not bolt it on. And building it in, truthfully it goes to that let's figure out your baseline. But then from a security profile, there's a lot of things that customers already have that they probably don't know they have. So Lenovo uses all the firmware information, any telemetry that's from Intel, it all goes into the systems that they have. And lots of times they're leveraging our trusted boot technology and other things, and the customers don't even know that that's happening, which I think is good. But then when they start to do and leverage AIOps, they sometimes buy things or add things on that are already there, that's already been built in, that's already stuff that the hardware vendors are providing for them, but they don't know it.

James Hilliard:

But they don't know how to say it.

Gretchen Stewart:

Exactly. Exactly.

Jay Bryant:

One thing I'd like to continue on the idea there is one of the aha moments early in AI that Lenovo had and that really spoke to me on the security side is the fact that you have different kinds of AI data that you want to keep segregated, and that's important. The information that employees need in their AI is going to be very different than what we need from an AIOps perspective. And so this idea of hybrid AI and the right information going to the right large-language models, the right databases and what have you, to keep people comfortable that their data's not going where it shouldn't be going.

Gretchen Stewart:

I think that's true, and there are companies like Intel who've built in secure enclave. So as an example, if you've got some anomaly detection product that you have and you're like, well, that's not something we've seen before, so you don't want that to be perturbating your entire environment. So you want to take whatever that bad actor is and put them in an enclave by themselves, in their own little jail. And Intel builds those capabilities, again in the intrinsics of the hardware that we build, those that folks like Lenovo and others take advantage of, but not everybody does.

And also other chip vendors that you work with, not Intel, don't have those same kind of capabilities. So folks need to really understand, and you can have, which most people do a mixed environment, some Intel, some AMD. You've got Nvidia. You've got a number of other things, but you also need to really understand what you're going to be able to get from some. And to your point about that secure data, maybe that shouldn't be on this set of equipment.

Jay Bryant:

Exactly.

Gretchen Stewart:

Or to secure this particular HIPAA data because, or financial data has got to be over here on these systems that have the capabilities that you need from a security perspective.

James Hilliard:

Every organization does have, we're no longer just a one type of technology shop anymore. That was the security strategy many decades ago. You know what? We'll simplify it by just all going in one area. Well, then all of a sudden, one area got compromised so then we started expanding, getting complex again. Okay, but now we have all these different systems out there. That does create a much wider security threat vector, and that's something that's got to be the leading conversation that you have in the data center these days of how do you protect all of it?

Chi Chung:

Yeah, I mean, we're used to a heterogeneous data center. With everything that's going on in the hypervisor world right now, we're now seeing data centers with multiple hypervisors. It's just something we're going to have to live with. But the reality is our clients, we have to make them aware that there's ways to simplify the management and that they need to embrace AI using, implementing AIOps to take a lot of the guesswork, a lot of the remedial stuff off their plates so that ultimately you can have an IT organization that focuses on proactivity, driving, innovating, bring innovation to the business, being able to work, take your limited resources and apply them towards higher value projects. That's the goal, and that's what we strive to work with customers on every day really.

James Hilliard:

AIOps to edge management, we've talked about the AIOps part of it. We haven't touched a ton on the edge, so I want to do that. Bismarck had a question in here, and this is right up your alley, Jay, "How does edge computing impact the data centers today?"

Jay Bryant:

That's a great question, and the reason I want to talk about that is we've really been focusing on edge. And with AI we think about data center, the need for large systems to process that, but you got to do the inference too. And in order to do that quickly and get that information, you need the systems close to the end user. And one of the, you see this proliferation of devices now, and the thought was, oh, well, it's at the edge now.

Well, the edge might be a smaller data center or it might be a system running under a conveyor belt, don't know, but it hasn't caused the data center to get simpler or smaller. It's really just made it have fingers out further into the world, which brings brand new considerations, the security aspect of it, bandwidth. Not every fast food restaurant has a T1 line into it or a gigabit ethernet. I'm amazed by the number. They're like, "Oh yeah, we still have a two meg connection." So you need to have those systems at the edge to do the inferencing, get the data ready, and then send it back to the data center where you have the larger systems to do the processing. So-

James Hilliard:

And you need that bandwidth so that you get those inferences made and back to make the decisions?

Jay Bryant:

Yeah, exactly. The low latency, all of that. So it's really just added another level of for us to solve. And I think, based on what I'm seeing with our customers, we've been talking edge, edge, edge for years now. I mean five, six years, but I think we're just starting to get to the point where customers are going to start seeing the pain points at the edge. It's something I've been thinking about for a number of years, and now the customers are starting to come going, "What are we going to do about this?" So it's another fascinating challenge.

James Hilliard:

Chi, you were nodding your head over there. Are those conversations starting to come in the Connection door?

Chi Chung:

Absolutely. And there's that, and then we still have a lot of customers who feel like, hey, my little PC at the edge is doing the job, and the reality is you're missing out by not investing in your infrastructure. When you sit down and you ask them, "Well, how are you managing your systems?" They go, "What do you mean? We just leave it out there and it just sits out there." It's not getting patched, and it becomes a security nightmare. So then when we go talk to the security folks, that's all we talk about is protecting that edge. And what you talked about is, I mean that is really we're at the beginning forefront of dealing with some of the edge issues because they're targets out there.

Jay Bryant:

And I mean at Lenovo, we've done some really cool things with the edge hardware from physical security and making it be able to survive if it's installed in a truck or on a ship or something like that.

Gretchen Stewart:

Exactly.

Jay Bryant:

But there's the software security aspect of it too, and the management, and those are the challenges that we're really going to next need to face and solve when it comes to that sort of thing.

James Hilliard:

And likely the challenges that you're facing out there now. Didn't get to all the questions that had come in, and we really appreciate everyone taking time to submit questions. There's one more question from Thomas I want to get to, and then I want to do a little bit of crystal ball action here, what are we going to be seeing moving down the line?

But last user question, Thomas had this one, and Gretchen, I've been waiting to ask you this one. It goes back to the AI side of things. "How do we ensure ethical AI development?" And that is close to you and something you spend a lot of time thinking and talking about. You and I could probably go for another two hours, but we can't.

Gretchen Stewart:

Absolutely, absolutely, absolutely.

James Hilliard:

But your high level thoughts on it.

Gretchen Stewart:

Well, the first thing is create a council. I sit on the responsible AI council at Intel, and literally we review model cards. We look at who's part of the information that's been building this, where do we get the data sets, are these okay data sets? I mean, you really have to have, again it's a team sport. So when you are working with developers, you've got to, as part of your sprints do that ethical review, have you done bias checks? Have you done some of these different types of testing that you really should be able to do as you are building it to ensure that, again, it's built in, not bolted on?

James Hilliard:

Right, absolutely. And something focusing on AI is something that the Connection team through the Helix Center focuses a lot on too. These conversations about ethics are coming up in that arena?

Chi Chung:

Yeah, definitely. I mean, I think we're just starting to have these discussions. And a lot of it is from the leaders up top because they're hearing, they're going to conferences, they're hearing things, and they're coming back-

James Hilliard:

Sitting on boards.

Chi Chung:

They're coming back to IT team saying, "Hey, we need this." And now the IT team is being tasked with if I decide if one, is it the right thing to do, and two, how do we do it with everything else we're doing?

James Hilliard:

All right, folks. So that's it for the user questions. I have to thank all of you for submitting those to us. It was great to see what was on your mind. We've covered a lot of ground here. We've talked about obviously there is some complexity in our data centers. There are some ways that we can simplify things. Strategy is a big way. Tools obviously, and then partnership, working with your teams and your ecosystem out there.

Let's talk about the future then. We've covered some ground to where we are. What do you want teams and people out there that are listening to think about in the coming months about their data center and how they can continue a simplification journey?

Gretchen Stewart:

I'm going to give it two things. One, I'm going to go back to something that Chi talked about, which is workforce.

James Hilliard:

Okay.

Gretchen Stewart:

I think it's really important to have your data center folks continually learning, but I also think looking at certifications, things like that, that can really bring people up to speed faster and train each other. But the other I'll call future sort of thing is we're working on AI right now, but it's really more humans are inputting, we're getting data and the AI is responding based on training, based on information that relates to predictions and things like that. But the next level is really self-learning where the chips themselves are self-learning. And at Intel we have something called, it's our neuromorphic computing product, and it is self-learning. So it is leveraging, think of your brain, and it's designed as your brain is in terms of the integration of the cortex, the other components in a memory CPU connection. As opposed to today, everything we have is CPU, memory, input, output. This is a fully integrated fabric that's memory and CPU.

James Hilliard:

Cool things to come.

Gretchen Stewart:

Very cool things to come.

James Hilliard:

Very cool. What's to come? What's on your mind?

Jay Bryant:

So I don't want to steal Chi's thunder, but I think the next steps for people are the first to break down those silos, start communicating with each other, educating each other. IT changes so fast and there is so much going on that you can't do it by yourself. I love the idea that it's a team sport. So start looking internally to yourselves first on what do we need to do to make this more efficient and easier for us?

And then second, reach out to your partners. Reach out to Lenovo. Tell us the problems you have so we can help you solve them. I want to do facial recognition at the edge. Okay, here's the things that you need to do that. We're starting to build those reference platforms, the examples to do it so that you don't have to do it yourself, and then you can be more efficient and keep your bosses happy.

James Hilliard:

Always important.

Jay Bryant:

Keep your internal customers happy and do really cool things, enable yourself to make the next, whatever it may be, the next app, the next cool thing that will be the shiny thing that we go after and go, "Ooh, I need that."

James Hilliard:

Or the thing that makes it more complex again down the road that then we'll have to simplify again in 10 years.

Jay Bryant:

We'll be back here talking about this again, but we'll be coming with even more information and more experience. So, you know?

James Hilliard:

That's ultimately the goal. That learning, and I appreciate you saying that, that learning from each other. Chi, something to grow on here. What are you looking forward to?

Chi Chung:

I would say I would challenge IT organizations to get a seat at the table with the stakeholders. Have influence, because if you don't, you're going to be at the back end of all these requests, and you're going to be responsible for trying to make things happen, right or wrong.

In addition to that, I would say that there's a lot of resources around that's available. The key is to stay up-to-date, stay relevant as to what's going on in the industry because it's going way too fast. Even for our team, we're constantly getting certifications, training, weekly, monthly basis. But all those resources are available to our clients, and they don't have to do this alone, and it doesn't have to be complex. I hope that's the takeaway for them.

James Hilliard:

One of the places you can find those resources, of course, is at Connection.com. And then if you do have your Connection account management team in place, they'd love to hear from you. You can learn more obviously about the Connection services, but also what the partners are doing at Lenovo and Intel and elsewhere.

So really appreciate you all taking time to join us for this Ask the Experts. Again, my name is James Hilliard. Thank you for joining us all. On behalf of Gretchen and Jay and Chi, we thank you for being here, and we look forward to talking to you all down the road.