A quarterly review declares VMware migration a success. The licence bill is down, the final cluster moved on time, and the programme lead is gaining commendations. Then along comes the auditor to look at the security configuration of the new platform and discovers 15 years' worth of exceptions copied over line for line. It seems that none of them has been looked at by the security team at all. The security team was only brought in to the project after the hypervisor had already been selected. Nothing has overspent, nothing has failed; the old risk has simply been relocated.
In this episode of The Security Strategist podcast, host Alejandro Leal and Analyst at KuppingerCole speaks to Sam Marland, Senior Manager and Virtualisation Sales Leader for EMEA at Red Hat, about virtualisation and what happens to security when an enterprise moves away from its virtualisation platform.
Cost is typically the reason for such a stray. The risk that Marland emphasises in the conversation with Leal is who is in the room when the new platform is selected and who then takes on the outcome.
As head of virtualisation sales at the world's leading provider of enterprise open-source software solutions, Marland also puts up a compelling case for OpenShift. Red Hat OpenShift is an enterprise-ready, Kubernetes-based application platform used to build, modernise, deploy, and scale containerised and virtualised applications across hybrid clouds, public clouds, and edge environments.
“They look at disaster recovery, high availability, and resilience. All those key topics come up very early in the decision lifecycle. Security less so,” Marland says, alluding to individuals on the lookout for a virtualisation platform.
Also Watch: Can Open Source Ensure AI Works For Everyone, Not Just The Largest Enterprises?
Why Cost is Taking Priority Over Security?
Customers are now moving away from VMware (now Broadcom). “The main reason people are moving is that the price has changed with their existing provider,” Marland says.
To what extent the price changed is a matter of dispute. CISPE and ECCO both claimed in May 2025 that VMware licensing had gone up to between eight and 15 times what it had been previously. A January 2026 survey of 302 North American IT decision makers by Cloudbolt showed 86 per cent of respondents reducing their use of VMware. The wholesale exodus is not happening, but there is certainly a move towards diversification.
For a cybersecurity team, the cost comes at the end. Disaster recovery, high availability and resilience are signed off first, so security inherits a platform whose architecture is already chosen and has to fit its controls around it. “We find ourselves trying to retrofit the security policies to the platform once it’s already been established,” Marland says.
The KuppingerCole cybersecurity analyst says that if you simply recreate the old setup on the new hypervisor, you're taking fifteen to twenty years of technical debt with you.
Marland then explains the difference between the old setup and a rebuild. For the customer who just wants to go from one point-and-click platform to another, there are options. However, for the customer who wants to modernise, the cost pressure is such that they're being pushed to complete their migrations and get rid of the old estate (at other times they may have been perfectly happy to run both platforms side by side).
Also Watch: How HashiCorp and Red Hat are preparing enterprises for AI at scale
‘An increased attack vector’
While price is the first pressure on a migration, hybrid is the second. Leal states that hybrid and multi-environment is the default rather than a stepping stone. Marland agrees and says:
Open Source and the AI Balance
How open source, data quality and small models are redefining who can build compliant, trustworthy AI beyond the largest enterprises.
“As you add hybrid to the equation for security teams, what that means to me is it is an increased attack vector, an increased footprint for vulnerabilities to be found.”
He ties this to the rate of discovery. “We don’t have to talk too much about Mythos and frontier security models. The number of vulnerabilities being found is escalating quickly.” That leans towards frontier AI models such as Anthropic’s Claude Mythos Preview, currently limited to a small group of enterprises under Project Glasswing.
Marland also talks about a rebalancing. Five years ago, public cloud was the panacea. Now, in large financial services and life sciences organisations, you see a more even split between on-prem and public cloud. Customers are building dual vendor strategies to reduce their exposure today. The shared problem is managing the VMs in a consistent way across all of it with a very limited number of skilled people.
Why Red Hat OpenShift Virtualisation?
Marland’s puts up a technical case for Red Hat OpenShift Virtualisation, which runs VMs alongside containers on the same Kubernetes platform. According to Red Hat, it’s included in all self-managed OpenShift editions and available on AWS, Microsoft Azure, Google Cloud, IBM Cloud and Oracle Cloud, with a separate Virtualisation Engine edition for enterprises that only want to run VMs.
The security argument is based on declarative infrastructure. “We are trying to limit the number of snowflake platforms,” Marland says. “The idea is that you define your perfect platform in code. And then when you deploy a new cluster, you are just pointing a pipeline and building that infrastructure.”
When the definition of good changes, he says: “I make that change in code. It’s peer-reviewed by a security person in this case. And then it’s rolled out across the fleet.” Changes start on a small number of clusters, are validated, then spread.
Leal adds that using Red Hat’s approach, an enterprise ends up with one policy model, one audit trail and one automation pipeline, which in turn means far fewer gaps between the teams and far fewer opportunities for misconfiguration.
However, Marland points to a staffing problem. A platform team can harden the whole stack, but he says the people who have run VMs for 15 or 20 years should stay as specialists, because the workloads, boundaries and interaction points differ from those of containers.
A Magic Quadrant Post a Decade
The episode comes weeks after Gartner published its first server virtualisation ranking in 10 years. The last one (the August 2016 Magic Quadrant for x86 Server Virtualisation Infrastructure) had Red Hat Enterprise Virtualisation on its own in the Visionaries quadrant. The 14th September 2026 edition kicks off with a very straightforward summary: the market is undergoing its most disruptive period in over a decade. It looks at 17 vendors, from Broadcom (VMware) and Microsoft to Proxmox, SUSE and Scale Computing.
According to the Virtualisation Review, Nutanix is at the top of the Ability to Execute chart, Broadcom is way out on the right in the Completeness of Vision category, and Microsoft and Red Hat are with them in the Leaders quadrant. Gartner says the renewed competition is down to Broadcom buying VMware.
Marlande says, "We are probably the most right-wing on the completeness of vision". He points to customers actually seeing the combination of virtualisation, containers and AI in a hybrid environment as proof of this. Red Hat has been in the business “seriously for about three years.” He anticipates that the company will make progress on that axis in the next edition.
When asked about a key takeaway for CISOs, Marland says to pick a platform that keeps your options open. Use the move to reset what good looks like. And put someone from the CISO’s team in the room from the start. “Don’t just make this a commercial point and a run to the finish line.”
Price gets a VMware exit approved, and the deadline it sets is why security arrives late. Marland argues the move is a rare chance to replace drifting, hand-built configuration with a platform defined in code and reviewed by security.
Key Takeaways
- Price drives most VMware exits; architecture rarely makes the shortlist.
- Security is usually invited after the platform is chosen.
- Retrofitting policy onto a live platform is the costly path.
- Like-for-like swaps carry 15 to 20 years of technical debt.
- Define the platform in code and deploy clusters from a pipeline.
- Config changes get peer review by security, then fleet-wide rollout.
- One platform means one policy model, audit trail and pipeline.
- Keep virtualisation specialists as a capability on top of the platform.
- Gartner’s first server virtualisation Magic Quadrant in a decade names Red Hat a Leader.
- Nutanix scores highest on execution; Broadcom sits furthest right on vision.
- Marland says open source caps vendor pricing through “fair market value.”
- CISOs should join migration discussions from day one.
Comments ( 0 )