The VMware Migration Math — Why the Real Cost Is the Data, Not the VMs

Hey guys, thanks for the intro. We're here with Susan Odell, who's the CEO of StorMagic, and we're having a little chat about, well, what goes into a VMware migration. There's a lot of interest, but it's not so easy.

Susan, welcome to the show. Thanks, Mike. It's a pleasure to be here.

Converting the virtual machine itself is relatively trivial, but I think where the challenge is, is that when I have a lot of data that I got to migrate, then it suddenly becomes a lot more challenging to deal with. And I think organizations are struggling with that aspect, and maybe that's why there's not as much of this migration effort going on, because people are weighing the cost of the migration versus the actual increase in licensing fees and coming up with some interesting math. But- Yeah ...

what's your sense of what's going on here? Yeah. It is extremely complicated, and it's not even just to your point, it's not just about migrating the VMs.

You've got even external issues to that. Hardware costs related to the infrastructure migration, supply chain issues. It's very complicated.

And the inclusion or the mandatory requirement now to move to VCF, there is no standalone vSphere where you can think about decoupling your storage layer and how you're managing that and licensing that. Puts a real conundrum out there in the market for anyone who has a significant fleet of servers that they need to deploy in terms of virtualization. Right?

So if you've got a single site or maybe you've got two sites, you can absorb the cost of having to license VCF now and then continue to operate business as usual. But if you get into high numbers of distributed edge locations, that mandatory requirement to VCF puts a significant burden from a licensing point of view. So I think your point around the economics of can I stay with VMware and move up to VCF versus the economics of scale, you've got to hit a pretty big number, in terms of distributed locations, for the economics to make sense to stay.

" It's another hard cost, but the cost of moving those VMs is a major project, and that is body and time for people to do that. Even if you don't have to look at new hardware, even if you're going to do an in-place migration, that's a significant project. So, the economics of making a decision are very complicated on top of the global economic and supply chain issues that are compounding those decisions.

And to your point, though, there are lots of organizations that had already opted for some other type of storage or networking solution, and if they go with VCF, those wind up either being licenses they don't use and are paying for, or they actually do the migration and that in itself is a significant cost, right? Yeah, absolutely. " As long as you can support HA, Broadcom is supporting vendors that still support high availability at the storage layer.

So we have customers today that are still running VMware, and they're running SV SAN as a storage solution because we do provide two-node high availability, and Broadcom is fully endorsing that kind of a configuration. So there is a decoupling of the licensing requirement from Broadcom from how you're deploying your infrastructure stack. So in some cases, customers may be in a situation where they're, for a period of time, having to pay for both things, and that might be okay.

In other situations, the burden of the licensing cost is less than the financial burden of moving, and they can keep technologies like StorMagic SV SAN in place, continue to run VCF, but only the hypervisor piece, and just keep operating business as usual. So StorMagic's in a pretty good position because we can offer a full HCI alternative to VCF, or we can help continue to support a diversified vSphere in the Broadcom ecosystem with SV SAN or HyperV or Proxmox in an alternative universe. What's your sense of what are people trying to do if they do decide to migrate?

Are they taking their applications as is in some sort of monolithic format, or are they attempting to maybe refactor some of them into a more modern platform and maybe move them into something that feels like Kubernetes or whatever it might be? Right. Yeah.

So my position has been, and will continue to remain, customers should be looking at their workloads first. What is it that they're running and where is the most appropriate place to run it? So certainly the hybrid cloud world environment is something we fully support.

If a workload makes sense in the cloud and if that consistent internet connectivity is available for certain workloads in certain locations, then you should think about that in that context. If it's a workload that needs to operate regardless of latency, regardless of internet availability, then it should run at the edge. And then you've got, okay, workloads, should it be in a bare metal container?

Should it be in a VM? How should we be deploying that workload? The reality is there's not a mass influx of workloads operating in bare metal containers.

We do support workloads in containers in VMs today. But that's not the major push at the edge, which is our space. So we are a distributed enterprise edgerobo SMB edge specific purpose-built technology company.

And in those environments, containers are not the driving force right now. Certainly, the preoccupation with VMware migration is the driver number one, and then modernization as economically as you can would be a second priority. And certainly, containers are not top of mind in that context today.

To your point about all of that, we live in the age of AI, and I keep wondering at some point, is somebody going to make it easier to migrate because they're going to come up with an AI tool to help me with the, I guess for lack of a better phrase, reverse engineering that might be required to do the migration? Well, the tools are out there today. We have a import-export tool, and we have a warm VMware import utility, which effectively minimizes almost all of the downtime from when you initiate a migration of a VM to when you enable it on our platform.

And so in the interim, production operations continue seamlessly until you cut over. In the enterprise data center, Nutanix has a similar capability that minimizes downtime. So I don't think AI companies are solving that migration problem today with the tech that we've got.

I think AI really from an edge computing context is really about being AI-ready, and that's about making sure that we can support the infrastructure that will support the workloads that will be running at the edge. And so that's something that's a priority for us in terms of our development focus, but not the AI capability itself. I've said in another discussion that I had a couple of weeks ago, infrastructure companies can't even begin to compete with AI companies Mm-hmm In terms of pace of how AI is going to develop.

So what we need to do is just provide the environment that lets those AI capabilities through partnership, through the ecosystem run, and let the AI companies worry about the latest and greatest AI code. Does this migration create an opportunity to maybe have a long-overdue conversation about good housekeeping when it comes to data? Because maybe we have a massive amount of data, but not all of it needs to move, and we can get smarter about that.

100%. I think the disservice, and I don't mean this in a negative way, but AI is like this nirvana of efficiency, right, and productivity. But if you don't have sound data at source, you're not going to benefit from it.

And I think I've been in enterprise tech for a long time, and so most of us, Mike, you'd be familiar with Gartner's Data Maturity Model as sort of the golden standard for how you move from your base state to prescriptive analytics. And most companies are not in a predictive or prescriptive state. And so I do agree with you 100%.

If people do the housekeeping and groom their data back to the data that is trusted and can be used to build basic, even just reporting functions of current state, what do I have today? Where am I today? Then you can start to build on that as you go forward.

So yes, I think that the global economy that we're in right now, which has created an environment of more pragmatic funding for projects, is an opportunity to reset, look at your data, and plan from a foundation that makes sense, and that starts with good data to begin with. Speaking of resetting, historically, we seem to, for whatever reason, separated the management of data and the underlying storage systems. Will that come back together?

Will there be more unification of data and storage management? Maybe we hear this phrase data engineers, but is the way we think about managing data changing? I just would repeat my last answer.

I think collectively, everybody is becoming super pragmatic about the infrastructure, the data and infrastructure, and management stack, all the way from what the system admins are doing on the top of the stack right down to the infrastructure and the data underneath it. There's a little bandwidth in every organization. So whether it is our end customer, IT organizations, our channel partners, our ecosystem partners, everyone is suffering from a bandwidth constraint, vendors such as ourselves.

So to me, that equals pragmatism. And so yes, I think that that relationship between data storage strategies, how that data is served up into an environment and serving the business, I think is just generally becoming more pragmatic and people are getting sharper around how that all needs to converge together to benefit the business in a more efficient way with limited resources and a lot of pressure. Will we unify the management of all these things then?

Because historically, we've kind of managed storage and networking and servers somewhat separately. We had some hyperconvergence around servers and storage, but is all that going to become a much more unified motion? I think that depends on the end customer and their strategy, right?

I think what we are seeing in some of our large enterprise customers, we certainly are seeing teams consolidateSo that you don't have fractional or siloed views of solutions. But I think that that's a business-by-business decision, right, in terms of how they want to function and operate. And I think that, given the Broadcom VMware example, having a unified stack 100% from a vendor in a particular area, I think there's some, it's not risk, but I think the consideration around that.

So I don't see in time that you're going to have a trend towards everything being unified in one way, but I think it is a company-by-company decision to say: How do we want to treat our IT strategy overall? And make decisions based on that. And as I understand the trade-off, it basically comes down to, yes, I can manage all of that in one motion, but then my upgrade cycles are tied to that.

And as far as I can tell, the technology advances are moving faster in one sector than another at any given time, so I don't want to be put in a position where I have to wait for the other two to come along and be compatible with the third innovation, right? So if there is an- Yeah ... advance in storage, I can move forward, right?

Correct. Yeah, that's why diversified stack, and the way that I talk about it is optionality. And the way that we look at things at StorMagic from a long-term point of view is we want to provide the market with the most optionality possible when it comes to edge-specific virtualization.

How can we give our customers the ability to say, "I want to work with StorMagic because they've been around for 20 years. Their tech is really solid. It doesn't go down.

I believe in diversification, so I'm going to go with StorMagic Storage," right? In another scenario, it may be, "I really appreciate full hyperconverged infrastructure because now I don't have to worry about the stack, and it's built on SV SAN, which is reliable. " That's the way we think about things, and I think it's really important for customers to make the distinction between the data center, the enterprise data center, and the edge.

They're not the same thing at all, right? In the enterprise data center, internet connectivity is not going to be an issue. You've got IT staff that's right there managing that data center, whether it's remote or literally on-site in a private data center.

They're managing it full time. At the edge, that's not the case. There's no IT staff.

You might have latency in one area is very different than latency in another area. And so I think that those distinctions of what are you managing, what's your strategy, do you believe in unified approach to different segments of your infrastructure, do you believe in diversification? And then look at your vendor landscape accordingly and see who's going to best support you, to your point, on the long-term journey, so as things change and modify, they can be more flexible and not have to rip out everything and start again.

So what's that one thing you see organizations doing today that just makes you shake your head a little bit and go, "Folks, I wish we could be a little bit smarter than that"? I think that I don't have a response to that. How I would respond to it is to say we're living in the most complex times in tech and in business ever.

So if I had any response, it would be anyone who's not slowing down to consider the times that we're living in, and to be planning pragmatically through that lens of chaos, right? I think everyone globally needs to be operating on a basis of managing change at lightning speed, and that no one knows where we're going to be in two years. And so how are you managing your environment to be able to absorb change that you can't see right now today?

Which is difficult. So that would be my response. If people are not operating through that lens of complexity and change, and am I functioning in a way that I can adapt, I can pivot, that would be a critical mistake right now.

So to your point about that, we hear the phrase "sovereignty" all the time now, and a lot of places and countries are thinking longer and harder about it, especially in the age of AI. As we go forward, though, ultimately, doesn't that just come down to having control over your data? It does.

At the base level, 100%, it's control over your data, but how are you hosting that data, right? So from a StorMagic point of view, sovereignty is fundamental to who we are because we are on-prem edge focus. It doesn't get more sovereign than that, right?

Mm-hmm. So yes, I would agree with that statement 100%. And this is why the edge computing space is growing so rapidly.

And we've been around for 20 years, so it's crazy. We've been purpose-built edge from day one, 20 years ago. It's our 20th anniversary this year.

And edge is scaling as an industry right now, and it gets tethered to sovereignty, data sovereignty. Absolutely. All right.

Well, folks, you heard it here. Hey, no matter what era we're in, it always seems to come back to the data. So how you manage it, it's the thing that makes all the difference.

Hey, Susan, thanks for being on the show. Thank you so much, Mike. Great conversation.

All right, and back to you guys in the studio.