When to choose vSphere and when to choose XCP-ng for VMware exits
If your estate depends on VMware-specific tooling and your budget can absorb per-core VVF or VCF subscriptions, vSphere is the lower-risk choice for now. If you want a Xen hypervisor with no license fee, managed through Xen Orchestra and backed by commercial support only where you choose to buy it, XCP-ng fits better. The decision turns on four things: how deeply your operations depend on VMware-specific components, how you design storage (vSAN versus XOSTOR), what support level your auditors and business owners require, and whether your team can run Xen and XAPI well. Start by listing every VMware dependency, including backup agents, monitoring integrations, automation scripts and third-party appliances, and check which of them vendors certify on XCP-ng.
Check your current renewal position before choosing. Broadcom's knowledge base states that ESXi hosts and vCenter on perpetual licenses keep running after Support and Subscription expires, but downloads of new patches and major version upgrades require an active support contract. That means an expired perpetual estate is not switched off, but it becomes harder to patch and upgrade safely. Confirm with your Broadcom account team which SnS dates apply to each host and whether your planned migration window falls before or after them, because that date sets how long you can wait.
Storage is often the decisive factor. vSAN users need a replacement for distributed storage, and XOSTOR is Vates' hyperconverged storage add-on, which Vates sells per host per year as a paid option. Ask Vates for XOSTOR's current requirements, supported hardware and the support tier that covers it, and compare that quote with the vSAN capacity you would otherwise need to replace. If you already run a SAN, check whether your XCP-ng storage connections meet your performance and multipath requirements before committing to a design.
Which platform fits which scenario
This table matches common decision scenarios to the platform that usually fits better, with the reason for each. Read each row as a conditional: the fit depends on your estate, skills and contract terms, so confirm the points in the last column against your own environment before deciding.
| Scenario | Better fit | Why |
|---|---|---|
| Renewal quote under VVF or VCF per-core pricing | XCP-ng with Xen Orchestra, if the quote is high relative to your budget for skills | XCP-ng has no hypervisor license fee, so the cost shifts to optional Vates support per host and internal operations. Get the vSphere quote in writing and model it against both options. |
| Heavy vSAN dependence | Stay on vSphere for now, or plan a Ceph-based move | XOSTOR is a Vates add-on, not a vSAN equivalent with the same feature set. Confirm which vSAN features your workloads depend on before choosing. |
| Heavy NSX dependence | Stay on vSphere for now, or evaluate OpenStack with Neutron and OVN | XCP-ng does not provide an NSX-style distributed firewall or overlay. Map each NSX policy to a target before committing to a migration wave. |
| No-license-fee preference with in-house Linux and Xen skills | XCP-ng with Xen Orchestra | The hypervisor and Xen Orchestra are open source, and the team can run patching and backups directly. Confirm that staff can operate XAPI pools and Xen Orchestra. |
| Need for paid SLA-backed support | XCP-ng with a Vates support plan, or a vendor-backed platform | Vates describes commercial subscriptions as adding SLAs, and Enterprise includes 24/7 support. Check the plan terms and the response commitments in writing. |
| Mixed estate with Hyper-V or XenServer history | XCP-ng | XCP-ng documents imports from Hyper-V and KVM and upgrades from XenServer and Citrix Hypervisor, so existing Xen-style tooling and habits carry over. Confirm which VM formats and guest tools your estate uses. |
| Team that wants one vendor for the whole stack | vSphere, or a single-vendor platform such as Nutanix | A single vendor simplifies support escalation, but XCP-ng support is split between the open-source community and Vates. Ask Vates which components each plan covers. |
| ESXi 6.x hosts still in service | Stay on vSphere for the migration window, or upgrade ESXi first | Current Xen Orchestra versions do not support ESXi 6.x for V2V, so those hosts need an upgrade or a different migration path first. Confirm the source ESXi version with your V2V test. |
How the Xen hypervisor differs from ESXi in daily operations
The main operational difference is architectural: XCP-ng is a type-1 hypervisor that boots on bare metal and combines the Xen hypervisor, the dom0 control domain, the XAPI toolstack and QEMU for guest I/O. ESXi also runs on bare metal, but its kernel and drivers are a VMware-controlled stack, while XCP-ng relies on a general Linux dom0 for host drivers and management. In practice, this means host driver and firmware checks are done against a Linux-based control domain, and troubleshooting often starts in dom0 logs and XAPI tasks rather than in vCenter alerts. Teams that already run Linux operations will find this familiar. Teams whose skills are almost entirely vSphere will need time to learn it.
The model matters most in three situations. Patching matters when you run many hosts, because XCP-ng updates are applied through the pool and XAPI, and you should confirm how rolling host evacuation and reboot sequencing work in your pool before you schedule waves. Hardware compatibility matters when you run older or unusual servers, NICs, RAID controllers or GPUs, because ESXi's hardware list is not the same as the Linux driver base that XCP-ng inherits. Guest drivers matter for Windows and legacy guests, because paravirtual drivers and guest tools behave differently under Xen and must be verified on each guest OS version you run.
Validate both points in your own lab before production. Install XCP-ng 8.3 LTS on the exact server models you plan to use, confirm storage and network controllers are detected, and run a full patch cycle with a rolling reboot of a test pool. Then migrate a representative set of Windows, Linux and appliance VMs with Xen Orchestra and check boot, network, backup and guest tools on each one. Record the results. Those records, not vendor claims, should decide whether the hardware and guest drivers you depend on are supported on this platform.
Where data sits and who controls the platform
With XCP-ng, the hypervisor, management plane and backups can all run in your own data center or in cloud accounts you control. The cards below cover where data and control points sit, how identity and audit evidence are handled, and which license terms your legal team should review before you modify anything.
Data location and hosting model
XCP-ng runs on bare metal in your data center, colocation space or a cloud account you manage, so VM disks, memory images and backups stay where you place them. Xen Orchestra can be deployed on its own management host, which means the control plane does not have to sit in a vendor cloud. Confirm where each component runs, including the Vates support tooling you may enable, and write the answer into your data flow diagram before the first VM moves.
Management plane access
Xen Orchestra is the web and API layer for XCP-ng pools, so its administrator accounts, API tokens and network reachability to hypervisor hosts define who can change VMs. Restrict the management interface to an administrative network, place it behind your own bastion or VPN, and log every API call to your SIEM. Vates subscriptions do not change who holds administrative control over your pools.
Identity integration
Confirm how Xen Orchestra authenticates users and whether it can take directory-backed sign-in through your identity provider. Map Xen Orchestra roles to your access model, separate operators from backup administrators, and keep break-glass accounts under documented custody. Because identity behavior depends on the installed Xen Orchestra version, verify SSO and role options against the current release notes in your own test pool.
Audit trails and compliance evidence
Audit evidence for XCP-ng comes from pool event logs, Xen Orchestra audit records, backup job reports and change tickets, and you must decide which of these feed your retention store. DeployOpen can plan the logging design, configure collection and prepare control evidence for an assessment. DeployOpen does not issue audit reports or certifications; independent auditors and certification bodies make those determinations.
Open-source licensing and internal changes
The XCP-ng repository is GPL-2.0 and Xen Orchestra is AGPL-3.0. Internal use without distributing modified binaries generally carries different obligations than offering modified Xen Orchestra as a network service, and the AGPL network clause is the term most teams need to read closely. Ask counsel to review whether your planned patches, scripts or forks trigger source disclosure, and document which components you modify.
Support and patch provenance
Under the Broadcom model, ESXi patches require an active contract. XCP-ng 8.3 LTS is supported until November 30, 2028, and you should confirm which update channel, signing and repository access your team uses and who approves patches. Record the source of every package in your change process so auditors can trace what runs in production.
XOSTOR versus vSAN and the support question
vSAN and XOSTOR solve the same problem, pooling local disks across hosts into shared storage, but they reach it from different places. vSAN is part of the vSphere stack, and Broadcom now sells it only inside VMware Cloud Foundation or VMware vSphere Foundation, so its cost sits in the per-core subscription. XOSTOR is an optional storage add-on for XCP-ng sold by Vates at $800 to $1,200 per host per year, according to the Vates pricing and support page. Each choice drives different design work. With vSAN, you size disk groups and storage policies inside vCenter. With XOSTOR, you decide how many hosts join the storage pool, how many disks each host contributes, which network carries replication traffic, and how a failed host or disk is rebuilt. Ask Vates to document the replication layer and its failure behavior in writing before you commit to a layout.
Storage sizing should start from a workload assessment, not from the host count in a quote. Measure the current vSAN or SAN environment over a representative period: IOPS, latency targets, read and write mix, capacity growth, snapshot and backup load, and how many hosts you must be able to lose before service degrades. Then run a pilot on a small set of representative VMs with the same disk layout you plan to buy. This also decides the order of work. Xen Orchestra's V2V tool can read disks from VMFS, NFS or vSAN datastores, so you can move workloads to XCP-ng before the new storage cluster is fully sized, but the storage that receives them still has to meet the measured requirements. Confirm the V2V source requirements against the XCP-ng migration documentation before scheduling waves.
Support is where free software and paid software diverge most. Vates describes the XCP-ng hypervisor and Xen Orchestra as open-source with no license fees, and says commercial support subscriptions add SLAs. Its VMS plans include Pro at $1,000 per host per year with a 3-host minimum and Enterprise at $1,800 per host per year with a 4-host minimum and 24/7 support, according to the same pricing page. Before signing, decide which tier matches your risk profile: how long a storage outage can last, who is on call, and whether a fault would stop regulated services. Then ask Vates for written response times by severity, whether the SLA covers the XOSTOR add-on, who handles a storage fault when the hardware vendor is involved, and how escalation works. Compare those commitments with the support already included in your VVF or VCF subscription.
How a move from vSphere to XCP-ng usually runs
The sequence turns a vSphere estate into XCP-ng pools with verified backups, tested guest drivers and rollback points for each wave, so the cutover happens on evidence rather than assumption.
Inventory VMs and their dependencies
Build a list of every virtual machine, its operating system, its disks and datastores, its network segments, its backup jobs and the applications that depend on it. The team that owns each workload signs off on the list, and the infrastructure team confirms which VMs use vSAN-backed storage or other features that need a different path. The output is a reviewed inventory with a proposed wave for each VM, and the list is the basis for every later step.
Confirm the ESXi version for V2V
Check the ESXi version on every source host against the Xen Orchestra V2V requirements in the XCP-ng migration documentation. The current V2V tool supports ESXi 7 and later and does not support ESXi 6.x. Hosts that fall outside that range need an upgrade or a different import method before they can be scheduled. Confirm the version with the documentation for the Xen Orchestra release you plan to deploy. The output is a host list marked as eligible or blocked.
Validate hardware and guest drivers in a pilot pool
Install XCP-ng on a small pilot pool built from the same server models that will carry production. Check CPU and firmware support, network and storage controller drivers, and the guest tools or drivers that Windows and Linux workloads need after conversion. Run representative application tests on migrated copies of non-critical VMs. Write down any hardware or guest issues and their fixes. The pilot is complete when the team signs off on the results and the recovery procedure for a failed migration.
Deploy Xen Orchestra and connect the pools
Deploy Xen Orchestra in your own environment, connect it to the pilot pool and the ESXi source hosts, and set up role-based access for the operators who will run the migration. Configure authentication and audit logging to match your existing identity and logging standards. The check is a successful connection to every pool and source host, with a test login from each operator account. Confirm the installation method and its updates in the Xen Orchestra documentation before you rely on it.
Run warm V2V migrations in waves
Migrate VMs in the waves defined in the inventory, starting with low-risk workloads. Each VM runs warm while its data replicates, then the tool syncs changed blocks after shutdown, which keeps the cutover window short. Before each wave, take a backup and record a rollback point, such as the source VM kept powered off rather than deleted. After each wave, the application owner checks boot, network reachability and service health. A wave is closed only when every VM in it passes those checks and the rollback point is documented.
Rebuild backup jobs and networking
Recreate backup schedules, retention rules and restore tests in Xen Orchestra for the migrated VMs, and map every network and VLAN to its XCP-ng equivalent. Verify that firewall rules, DNS records, load balancer targets and monitoring checks point to the new VM identities. Run a test restore of at least one VM per workload type. The output is a backup report showing successful jobs and a restore test that the application owner has accepted.
Cut over production and retire vSphere
Switch production traffic to the migrated VMs according to the cutover plan agreed for each wave, and keep the rollback point available until the application owner confirms stable operation. Once all waves are closed, record the final state of the estate and confirm that no workload depends on the old vCenter. Decide with the licensing team how the perpetual or subscription ESXi hosts are retired, and confirm the terms with Broadcom before removing support contracts. The cutover is complete when the inventory shows every VM running on XCP-ng.
Decommission ESXi and document the new estate
Remove the ESXi hosts, vCenter and any remaining vSAN disks from production, after securely wiping storage that held sensitive data. Update the configuration management database, runbooks, the backup policy and the patch procedure to reflect XCP-ng pools and Xen Orchestra. Store the migration records, the wave sign-offs and the restore test results with your audit evidence. The step is complete when the decommission record is signed and the operations team can run the new estate from the documentation alone.
Questions buyers ask before switching from vSphere to XCP-ng
What drives the cost of moving from vSphere to XCP-ng?
The main recurring cost you remove is the VMware subscription, because XCP-ng and Xen Orchestra have no license fee and Vates describes both as open-source projects. The costs that remain are one-time and recurring. One-time costs include planning, migration waves, rebuilt backup jobs and network mapping. Recurring costs include hardware, storage, staff time, backup software and any Vates support plan. Costs rise with the number of hosts and VMs, the storage design (XOSTOR is an optional add-on that Vates prices at $800 to $1,200 per host per year), and the number of applications that need a sign-off per wave. Ask any provider for a written scope that lists assumptions per host, the migration waves, the support tier and what is excluded.
How long does a wave-based move take for our estate?
The duration depends on the number of VMs, their sizes, the number of hosts, the storage design and how many change windows each application owner can accept. No single duration applies across estates. Xen Orchestra's V2V tool copies disks while the source VM keeps running and then syncs changed blocks after shutdown, so downtime per VM is mostly the final sync and first boot, not the full copy. A practical pattern is a pilot wave of low-risk VMs to test networking, drivers and backups, followed by production waves grouped by application. Ask for a timeline tied to your inventory in the written proposal, and check which waves depend on each other.
Can our internal team migrate to XCP-ng without outside help?
Many internal teams can do it if they already administer Linux, have a tested backup platform and can dedicate time to a pilot wave. Xen Orchestra handles V2V from supported ESXi versions over HTTP access to the ESXi API, so the tooling is not the hard part. The gaps are usually Xen pool design, shared storage, network mapping, Windows guest checks, rebuilding backup jobs and writing the cutover runbook. Outside help is most useful when the estate is large, when the team has not run a production cutover before, or when compliance evidence must be produced for the migration. DeployOpen plans, implements and supports open-source platforms in your own infrastructure under a fixed scope agreed in writing before work starts.
What skills and change windows does the migration require from us?
The work needs four things from your team:
- Administration skills for Xen pools, XAPI-based tooling, storage and networking on Linux
- Named owners for each application who can approve cutover windows and test results
- Change windows for each wave, including a rollback plan if a VM fails to boot
- Validated backups for every VM before it moves, with a restore test that proves the data can be recovered
Your security or compliance staff should also review how access, logging and patching change after the move. Ask the provider to document each step in the runbook so your team can repeat it for later waves without relying on one person.
Can ESXi 6.x hosts be moved to XCP-ng with V2V?
Not with the current Xen Orchestra V2V tool, because the XCP-ng migration documentation states that ESXi 6.x hosts are not supported for V2V. The supported path is V2V from ESXi 7 and later, which reads disks from VMFS, NFS or vSAN datastores. If your hosts run ESXi 6.x, the practical options are to upgrade the source hosts to a supported ESXi version before the move, or to export VMs through another route and import them. Confirm which import paths apply to your versions with the XCP-ng migration documentation and test them on a non-production VM before you plan any wave.
What does XCP-ng 8.3 LTS support mean for our planning?
XCP-ng 8.3 LTS was released on October 7, 2024 and is supported until November 30, 2028. Support for XCP-ng 8.2 LTS ended on September 16, 2025, so any host still on that branch falls outside the support window. Plan the migration so that production runs on a supported branch throughout the waves and for the years after cutover. Schedule the host installs and any upgrades before the first production wave, and check the XCP-ng release documentation for the update path and the support dates that apply to your versions. Build a calendar that leaves room for a second wave if the first one exposes problems.
What does a Vates support plan add over the free software?
The free software gives you the XCP-ng hypervisor and Xen Orchestra with no license fee, but your team handles troubleshooting and patch validation. Vates' commercial plans add support commitments, including SLAs, and the Enterprise plan includes 24/7 support. Vates lists Essential at $2,000 per year and Essential+ at $4,000 per year for up to 3 hosts, Pro at $1,000 per host per year with a 3-host minimum, and Enterprise at $1,800 per host per year with a 4-host minimum. XOSTOR is priced separately. Compare the plan terms against your own uptime needs, confirm which components each plan covers, and review the Vates pricing and support page before you request a quote.
Questions to answer before choosing either platform
Before comparing features, confirm the internal facts that decide between vSphere and XCP-ng. Start with VMware feature dependencies: list every vSphere capability your estate relies on, such as vMotion, HA, DRS, vSAN, NSX, Site Recovery Manager or Aria, and check which of them sit inside the VVF or VCF bundle you would be quoted. Note that Broadcom states many of these are no longer sold standalone, so a feature you use today may only be available inside a larger subscription. Also record your renewal dates for each perpetual or subscription contract, since Broadcom's knowledge base says perpetual hosts keep running after support expires but patch downloads and major upgrades need an active contract.
Next, document host count, CPU generation and hardware compatibility. XCP-ng runs on a shorter hardware list than vSphere, so confirm each server model, NIC and HBA against the XCP-ng hardware compatibility guidance before purchasing anything. Confirm your storage type as well: if you run vSAN, XCP-ng uses XOSTOR as its paid storage add-on and the storage design must be re-planned, while SAN or NFS datastores are easier to keep. The V2V tool reads VMFS, NFS or vSAN datastores, so verify which ESXi version each host runs, because the XCP-ng migration documentation says ESXi 6.x hosts are not supported for V2V. Check the ESXi version on every host, not only on the newest one. Then assess people. Count the engineers who can operate Linux, Xen, XAPI and Xen Orchestra backups, and decide whether training or outside help covers the gap. Compliance evidence is the third area: list the controls your assessor expects for change management, backup restore testing, patch history and access logging, and confirm the tooling each platform will produce for them. Finally, decide whether you need paid per-host support from Vates. Its published tiers add SLAs and 24/7 support on top of free software, so the answer depends on your internal on-call capacity and your contractual obligations, not on the platform alone.
How DeployOpen helps plan, implement and migrate XCP-ng
DeployOpen plans, implements, integrates, hardens, backs up and documents self-hosted open-source platforms inside your own infrastructure. For a move from vSphere to XCP-ng, we start from your written scope, so the V2V waves, backup jobs, networking and support plan are agreed before any VM moves.
How we scope
- We inventory ESXi host versions, VM count, guest operating systems and any vSAN-backed disks before proposing a plan.
- Engagements are fixed-scope, with deliverables, migration waves and timelines agreed in writing before work starts.
- We confirm whether Xen Orchestra V2V covers each source host version and which VMs need a different migration path.
How we build
- We deploy XCP-ng pools, Xen Orchestra and the storage design in your cloud accounts, data centers or air-gapped networks.
- We rebuild backup jobs, network mapping and VM settings, then test restores before any production cutover.
- We run warm migration waves with rollback points, checking guest boot, drivers and application behavior after each wave.
How we hand over
- You receive runbooks for host patching, pool operations, backup restores and the V2V migration steps used.
- We document the configuration, access model and decisions so your team can operate the environment without us.
- Handover includes a review of remaining ESXi workloads, so you can confirm what is still on perpetual hosts.
How we support
- Ongoing support, upgrades and continuous testing are available as retainers after handover.
- We can plan upgrade windows around the XCP-ng 8.3 LTS support period, which you should confirm against the release documentation.
- Where you choose a Vates support plan, we align our runbooks with the SLAs and support channels in that agreement.

