Why the enterprise virtualisation conversation changed

Broadcom acquired VMware and raised the prices. That’s the short version. The longer version involves per-core licensing changes, the elimination of perpetual licences, and a general attitude towards the existing customer base that can charitably be described as indifferent.

Overnight, organisations that had been running VMware for a decade or more found themselves facing renewal costs that simply didn’t make commercial sense. Some were mid-market companies running a couple of racks. Some were university departments. Some were defence organisations with tightly controlled budgets and long procurement cycles. None of them had budgeted for a threefold increase in virtualisation licensing.

What Broadcom actually did was force people to evaluate Proxmox properly – to run it up in a lab, test their workloads on it, and form a view based on what it could actually do rather than what they assumed it couldn’t. When they did that, most of them found it met the bar. Version 9.1 is the release where the remaining excuses ran out.

What “Enterprise-Class” actually means

The virtualisation industry let VMware define this term for twenty years. vCenter, vSAN, DRS, NSX – if your platform didn’t have direct equivalents, it wasn’t enterprise. That framing was always self-serving, positioning VMware’s product portfolio as the benchmark against which everything else was measured.

In practice, enterprise means something simpler than a feature matrix. It means the platform runs production workloads reliably, day in, day out, without surprises. It means you can back it up with tools your compliance team recognises and your auditors won’t question – because in regulated environments; “we use a community backup script” is not a conversation anyone wants to have. It means you can manage your infrastructure at scale without the tooling itself becoming a project that requires its own dedicated team. And it means that when something goes wrong at two in the morning, you can get someone competent on the phone who understands your environment and can help you fix it, not someone reading from a flowchart in a different timezone.

That’s the bar. It has always been the bar. The difference is that for twenty years, meeting that bar also meant accepting VMware’s commercial terms, because there wasn’t a credible alternative that ticked every box. Proxmox now clears it, and it does so without requiring you to remortgage your datacentre budget for the privilege.

What matured in Proxmox VE

Ceph Integration and Hyperconverged Storage

Ceph integration is what turns Proxmox from a hypervisor into a hyperconverged platform, and it’s the single biggest reason the enterprise conversation has changed. For deployments of four nodes and above, you get fully converged storage managed from the same interface as your VMs and containers. There is no separate SAN to procure, no SAN licensing to negotiate, no fibre channel fabric to design and maintain, and no separate storage team required to manage a parallel infrastructure. The storage is distributed across your compute nodes, it scales as you add capacity, and it’s resilient by design.

For the mid-market deployments we typically build – one to two racks at workgroup or department level – this represents a genuine architectural simplification. You are reducing the number of moving parts, reducing the number of vendor relationships, and reducing the number of things that can go wrong at three in the morning. Below four nodes, Proxmox supports shared storage over iSCSI, NFS, and Samba – which is what you need for HA clustering, since local storage alone won’t give you live migration or automated failover. These are well-understood protocols, they work with existing storage infrastructure, and they’re straightforward to configure. But once you cross that four-node threshold, Ceph gives you resilient distributed storage that grows with your compute estate, all managed from a single interface. That removes one of the most significant cost and complexity barriers in traditional virtualisation deployments.

It’s worth being clear that we’re talking about VM storage here – the infrastructure that underpins your virtualised workloads. For organisations with broader storage requirements – archival, high-performance storage for HPC workloads, research data – we architect comprehensive storage solutions alongside Proxmox, typically built around TrueNAS and purpose-designed for the specific use case. Ceph solves the hyperconverged compute storage problem. It’s not intended to be your entire storage estate.

Veeam Backup Support for Proxmox

For me, it was Veeam’s decision to officially support Proxmox that really moved the needle, and it matters for two distinct reasons:

The obvious one is practical – a huge number of organisations already run Veeam as their backup platform. It’s embedded in their processes, their compliance documentation, their disaster recovery runbooks. Proxmox support means that migrating away from VMware doesn’t require them to simultaneously rip out and replace their backup infrastructure, which would be a project in its own right. That barrier is now removed.

The less obvious reason is arguably more significant. Veeam is a company that makes commercial decisions based on where the market is heading. Their entire business model depends on supporting the platforms that their enterprise customers rely on. When they chose to invest engineering resources in building and maintaining official Proxmox support, that was a statement about where they see the virtualisation market going. For IT directors and procurement teams who need external validation before they’ll sign off on a platform change – and in enterprise environments, that’s most of them – the fact that Veeam have put their name to Proxmox carries real weight. It’s third-party validation from a vendor with no particular axe to grind.

Proxmox Datacenter Manager

The Datacenter Manager addresses what was, until recently, a legitimate gap. Historically, each Proxmox cluster was its own island – you managed it through its own web interface, updated it independently, and if you needed to move workloads between clusters that was a manual exercise. For a single deployment that’s perfectly workable, but for organisations running Proxmox across multiple sites or departments, the lack of centralised management was a genuine operational concern.

The Datacenter Manager provides cross-cluster visibility, cross-cluster migration, and estate-wide update coordination from a single point. It is not vCenter – vCenter has had twenty years of continuous development. But it solves the problems that actually matter at the scale we deploy: knowing what’s running where, moving workloads between clusters when you need to, and keeping your infrastructure patched without logging into each cluster individually. It’s a strong foundation, and it’s developing at pace.

High Availability Clustering

Proxmox’s HA works. Corosync for cluster communication, automated failover via the HA manager, proper fencing. When a node fails, the system detects it, fences the failed node, and restarts the affected workloads on surviving cluster members. The configuration is straightforward and it doesn’t try to do anything clever beyond what’s needed, which in infrastructure is generally a virtue.

Proxmox vs VMware: The DRS question

This comes up in every VMware migration conversation. DRS (Distributed Resource Scheduler) automatically rebalances VM workloads across hosts. Proxmox doesn’t have it. For a lot of people, this is where the evaluation stalls.

DRS is genuinely valuable at scale – hundreds of hosts, thousands of VMs, constantly shifting workloads. Nobody is disputing that. But most mid-market organisations are not operating at that scale. At one to two racks, you know your workloads. Proxmox provides affinity rules that give you control over placement – which VMs run together, which stay apart, which hosts are preferred. It requires you to think about placement rather than delegating it to an algorithm, but at this scale the overhead is minimal and the control is arguably better, because you’re making deliberate decisions rather than trusting a scheduler that doesn’t understand your business context.

If you genuinely need DRS, you probably need it because you’re operating at a scale where VMware is still the more appropriate platform. That’s fine. But conflating the absence of one feature with a platform not being enterprise-class is a category error.

Enterprise Proxmox Support in the UK

The standard objection to any open-source platform, and it’s a reasonable one. Organisations running production workloads need to know that support exists, that it’s competent, and that it’s accessible when they need it. “Read the documentation and post on the forum” is not an acceptable answer for a platform running your production infrastructure, and anyone who suggests otherwise has never had to explain an outage to a board.

Proxmox address this directly. They sell tiered enterprise support subscriptions – Community, Basic, Standard, and Premium – with increasing levels of response time and access.

We recommend Standard as a minimum for any production deployment. It provides direct access to the Proxmox support team with reasonable response times, and it includes access to the enterprise repository for tested and stable updates. This is not a hobbyist arrangement. It’s a commercial support relationship with SLAs and accountability.

On top of Proxmox’s own support, we provide an additional layer as the UK’s only Silver Partner. What that means in practice is a single point of contact with UK-hours phone support – you call a number and speak to someone who understands Proxmox and understands your environment. Not a ticket queue, not a chatbot, not a first-line script reader who escalates everything. We offer terms and credit for purchase orders, which matters more than people outside of procurement realise – a lot of organisations, particularly in the public sector, simply cannot process credit card payments to overseas vendors. Their procurement process requires a UK-based supplier who can issue invoices on agreed terms. That’s what we provide.

We are also bringing a 24/7 support service online for organisations that need around-the-clock coverage, because not every outage is considerate enough to happen during business hours. Defence environments, research facilities running time-sensitive workloads, organisations with global users – these need to know that someone will pick up the phone at any hour.

The support ecosystem around Proxmox is still developing. That’s a fair observation and there’s no point pretending otherwise. Two years ago the partner ecosystem in the UK was thin. But the direction of travel is clear. Companies like ours are investing seriously in building out the support infrastructure that enterprise customers require, because the demand is there and it’s growing. The gap is closing, and it’s closing quickly.

Who should consider Proxmox and who probably shouldn’t

Where Proxmox fits

• University departments running research computing. Proxmox is already widely used in UK academic virtualisation environments.
• Defence organisations building simulation environments within controlled budgets. Proxmox supports offline activation with offline licence keys, which matters when your infrastructure sits on an air-gapped network. Try doing that with VMware.
• Businesses running one to two racks who need something reliable, supportable, and cost-effective. At that scale, Proxmox 9.1 is not a compromise. It is a strong platform evaluated on its own terms.
• Public sector organisations with long procurement cycles who need a UK-based reseller who can issue invoices on agreed terms.

Where it doesn’t

• Hyperscale multi-datacentre estates. VMware remains the more mature option. DRS, NSX, vRealize, and the third-party ecosystem around vSphere exist for a reason at that scale. Nobody is suggesting a Fortune 500 company migrate its entire global estate to Proxmox.
• Secondary schools on an all-Microsoft stack. Hyper-V with existing Datacentre licensing is probably more pragmatic.

A Note on Cost

The support subscriptions are per-socket and increase modestly year on year. Nobody should pretend it’s free. But it meets the enterprise standard without aggressive licensing ratchets, and without the risk that the next corporate acquisition renders your infrastructure commercially unviable. For a growing number of organisations, that last point is the entire reason they’re making the change.

---

Novatech is the UK’s only Proxmox Silver Partner. We supply hardware, support subscriptions, and migration services – with UK-hours support, purchase order terms, and direct access to our engineering team.