IT Challenges of Multicloud and Hybrid Cloud Computing – Chad Dunn, Dell

This is texturing TV. Hey guys. Thanks for the throw.

We're here with Chad Dunn who's vice president of Apex management for Dell and we're talking about multi-cloud Computing and maybe hybrid cloud and the crazy old world. We live in today Chad. Welcome to show.

Hey Michael, thanks. It's great to be here. What exactly is going on?

I thought we were on this march to hybrid cloud computing and we were gonna have this one centralized console the manage everything and I was gonna get flying cars and all kinds of good things to go but seems like today we have workloads just about everywhere. You can imagine summer on premise somewhere on multiple clouds and maybe the total cost of it is getting a little out of control as a result. Where are we?

How did we get here? And how do we kind of bring some sanity back to the whole thing? Well, number one.

I was also promised to fly a car and I've yet to see it but you're right. You know, we we do find ourselves in a hybrid Cloud world if we look at our customer base 90% of them have hybrid cloud in place. But we're also in a multi-cloud.

So our customers are using multiple, you know public or hyperscaler clouds to run their workloads and a lot of times they they found themselves here. By default, right? So, you know, they've got lines of business and Engineering organizations who are the real Innovation engines within those companies and they're looking for best to breed services in you know, things like iot and Ai and ML and databases very often those preside in different Cloud providers.

They're gonna migrate to where those tool there their workloads are best to for those services. So at some point you you sort of wake up and realize I'm in a lot more clouds than I thought. I was a lot more work loans.

And maybe sometimes I've gone there maybe not for the right reasons, you know, sometimes we've seen a lot of sort of top-down mandates that we shall go to the cloud. And they've gone there and not really given thought to what workloads belong there. And what don't at some point you sort of realizing and you look at the you look at the bill and go.

Wow, you know, I'm I'm in a lot of places. I didn't necessarily know I was in and paying a lot more than I thought I was paying and that's not a great place to be right. We want to get customers to multi-cloud by Design rather than by default.

to your point Are people moving workloads once they discover that perhaps this workload the nature of it has changed and it's more expensive than running one place versus another or they just kind of once the workloads there. Everybody's like well, I don't know if I want to risk my career moving and workload around. So what is the appetite for movement?

Well, you look at like anything. It depends. It depends on what cost you're really bearing for that service.

So as equally delicate as moving a workload back is it's equally delicate to you know, face your your leadership with a cost and maybe out of control or at least significantly higher than then you expect or they expected. So, you know, there's some amount of career preservation balance in in that but we we did see a lot of things just run to the cloud and then after the fact realize wow, I'm paying a lot more and this is this is something I'm gonna have to correct and and I don't want to give you the impression that there's widespread repatriation and workloads out of the cloud that there's an exodus other club a lot of these workloads belong there that's exact right place for them. And I think this this repatriation is It's exception not the norm, but it's still significant.

what is the level of sensitivity around cost these days and I'm asking the question because the most expensive thing in the it budget is the people and it seems like every time we add a new platform we get a new console and we hire somebody to sit there in front of that console. So is that just mathematically insane and do we need to figure out a way to let people manage multiple platforms at scale without necessarily having a baby City each one? Well, I think what we're what we looked to provide the customer is a way to be able to service both of the needs of Having their data where they want it whether it's on-prem or whether it's in the cloud and seamless Mobility that would be part of our ground to Cloud strategy.

So putting our software IP into public cloud and also give them the ability to bring some of those public cloud services on premises. That's our Cloud to ground strategy where there are multi-cloud appliances allow you to run instances of of public Cloud on your premises for compliance and security and then give you that that ability to manage it all in context. So to be able to see where my data is and move it whether it's in public co-location or arm in my data center the same with those workloads, you know, are they running truly in the public Cloud?

Can I move them and run that operating model in a Cola? Can I run that on my premises and give them the the visibility the control the Telemetry to be able to make those make those decisions, you know multi-cloud as a fact and it's going to continue to be a fact for quite a long time because just not all those workloads are suited to be there and you know, we've certainly seen, you know cases where you know customers are you know bringing those workloads back what they do struggle with, you know, the number of single Pains of glass that they have to manage that they have to manage with but to some extent that's a fact of life and what we will we seek to do, you know through giving them the visibility As at least be able to look at and manage all of these different Cloud presence presences in context of one another. If you look at this from a X Y axis kind of thing, we've been talking about the y axis where it's the number of platforms, but on the X side is equation the types of applications that people are deploying is changing as well.

We've got Legacy monolithic stuff on version machines. We've got containers running on kubernetes and serverless Computing framework. So is that all so adding exponentially to the complexity?

Yeah, it absolutely is and a lot of this again what happened with these sort of top-down directives from the CIO to say, hey, we're gonna go to the cloud and you have to make an excuse why something shouldn't go to the cloud now that should be an opportunity for your companies to you know, rationalize and re-examine those workloads and say is this something that Simply needs to stay on-prem. Does it need to be refactored or re-platformed to go into the cloud or is it a lift and shift or is it something that could just simply be rationalized away? What's my actual usage on this application?

That's that's been around for the last you know, 10 15 years and I don't know the dependencies of it. So it shouldn't be an opportunity to do a little bit of spring cleaning very often that doesn't happen and you get sort of this on mass, you know migration of these these workloads to public Cloud but that's what should be happening and you've got to take a you know, if I were to give a company's the advice is do the hard work and do that rationalization that spring cleaning of those workloads before you undertake this because it's a lot easier to do before you've undertaken the cost and the effort of the refactoring or the re-platforming of those of those workloads, and we certainly see customers who have made the move. And then come back so that actually give you two examples.

So one customer of our 37 signals, they're actually moving a lot of their workload off of public cloud back on premises because it simply for the workloads that they were offering did not make sense to be there. Now. They they spent something around $600,000 worth of Dell servers to run these workloads.

But by bringing those back on premise they believe they're saving seven million over the next five years, right pretty dramatic example of a workloads that probably shouldn't have gone there in the first place that weren't refactored in order to be replatform. And then, you know companies that that go through that math and go you know what I shouldn't go there in the first place. So another customer of ours called href now, they they believe by their calculation they say 400 million over three years by not moving to the cloud now, like it's always tenuous just to say what the cost of avoidance was but that's an example of a company that did the hard work of that that spring cleaning process of saying hey do I really need this workload?

And what's it really gonna look like once I refactor it where I lift and shift it onto the cloud and that cost avoidance can be significant. This may be a gross over a simple vacation but it seems like the longer an application runs consistently in the less spiky. It is the better it is to run on premise because I'm aggregating the cost of running that across the infrastructure versus the spiky monthly bills that I might see in the cloud and I'm sure that a lot of people are going.

Well. I wish it was that simple but you know at the end of the day, that's a calculus then my mind works. Yeah, it's it's a generalization.

But I think it's largely a true generalization. You know, I don't want to fall into the the sunk cost fallacy. But you know, if you've got something that's that's running reliably and it is a relatively stable workload.

If there's a good there's a good chance that it belongs on premise of belongs right where it's at. But that's not the same for all work clothes. Now, you've got to take into account the other things of as these workloads, you know could potentially go into the cloud as as they have and not through really Shadow it but through you know lines of business looking for those services.

Their it organizations need of exercise some level of control and governance and stewardship and security over that data. And you know a lot of times they're going to feel a little bit out of control as these workloads start to sprawl into into public cloud. And so that's again.

I'll go back to our grounded Cloud strategy where we say, you can take our our data services IP our storage IP. Run that on premises run that in public Cloud run that in co-location. In fact run that in multiple public clouds get full data mobility and security with one operating model.

That's pretty powerful for that for an IT organization because then you can start to make those compliance and governance decisions knowing that you have full control. And that also gives you the opportunity to locate the data where it needs to be and not undergo some of these big egress costs. So if you have a very intense data operation it moved all your data into into you keep your data on Prem.

For compliance and security reasons, but you run the workload in the cloud that we've seen companies do this and then get that First Billing. Oh, wow. Look at the egress cost of that data when I'm operating on it on Prem and then bring bringing results back.

It can be pretty significant. So it really comes down to you know, intelligent workload placement intelligent data placement and giving the customer the visibility to do so now Apex in my mind at least is a set of managed Services provided by you guys to help deal with all this complexity. Back in the day, I think maybe.

All of it consumption maybe 20% was consumed via managed Services. Is that changing? Are we seeing organizations kind of having a different mindset when confronting with all this complexity and more people shifting towards consuming IT services.

They are shifting to consuming as a service but they're all so shifting to Simply consume via subscription or other flexible consumption models. So I would separate the financial construct of how they consume from whether or not the customer or customers are managing it themselves. So we we see the transition to say subscription and other flexible consumption models happening more quickly, but you know, we're still seeing significant move toward as a service say vendor manage down managed in in our case, but certainly the financial construct is moving a little more quickly than you know full as a service changeovers.

Now one thing that we have noticed is when it's something that's a very novel or net new service and I'll take for example our cyber Recovery Solution in in Del Apex Where there's not an already an existing competency the center of Competency within the customer account on doing a certain thing like cyber recovery and ransomware protection much more apt to be able to consume that as a service rather than to build out that competency. within within their own environment and probably less so opportunity to take out something that has a service when there's already a rich set of skills already in that in that customer's environment. So cyber recovery may be much more popular to consume as a Dell managed service versus Simply Storage.

We would typically see more customers take advantage of subscription and flexible consumption models on storage then us managing it because they have those people they have those skills. They've been doing it for a while. Do you think the relationship is changing maybe along this line as well managed Services was kind of an All or Nothing proposition, but it feels like we're moving to the middle and these are really co-managed.

Right the customer has their it people involved you're involved. They may be a little more weighted towards the applications and you're more towards the infrastructure, but the Dynamics are changing. Yeah, I think that's right.

And I think that's that's a function of customers looking at where they're spending their money and the more sort of well-known and well understood and you know there I say mundane tasks of infrastructure management as as good capitalists. We push things left in the supply chain. Once they become, you know, pretty normalized and and pretty simple to do and so we see that so things like, you know life cycle management and upgrades and updates and things like that.

That's that stuff is significantly easier for customers to push in and supply chain and consume as a service versus some of the more interesting things. They may want to do with it and more and more. We're seeing those interesting things be developer LED and we're saying much more influence of the developers within our accounts, you know, having much more influence in the decision making process around infrastructure around services and the way that they'll consume All right, folks.

Well, you heard it here. You might want to come up with a list of all those devops things and put them in two categories. The ones that are fun and the ones that are not fun take the list.

There you go find and call Chad. Hey Chad. Thanks, you know the show.

All right, Michael. Thanks. I really enjoyed it.

All right back to you guys in the studio.