Virtualization and VMware

Oracle on Nutanix and other platforms.

Oracle on Nutanix is treated under Oracle's partitioning policy the same way as VMware, which means the policy does not recognise it as hard partitioning and a cluster wide claim can follow. Because that policy is not the contract, these findings are often cut 60 to 80 percent once they are read line by line against the signed agreement.

How does Oracle treat Nutanix for licensing?

Oracle treats Nutanix the same way it treats VMware, because its partitioning policy does not recognise the Nutanix AHV hypervisor as hard partitioning. In Oracle's published view, a hypervisor that lets a virtual machine move between hosts cannot reduce the licensable footprint, so the requirement extends to every host the Oracle workload could run on. On a Nutanix cluster that can mean licensing all nodes, not just the nodes where the database currently sits. This is the same reasoning that produces the large VMware findings, applied to a different platform, and it lands the same way: a preliminary number inflated at list price.

It helps to be precise about what Oracle is and is not saying. Oracle is not claiming that Nutanix is technically unable to constrain a workload. It is claiming that its own policy does not accept that constraint as hard partitioning. That is a policy position, not a fact about the technology, and it is also not, by itself, a term of your contract. The same distinction that governs VMware governs Nutanix, which is why the defense starts with the agreement rather than with an argument about the platform.

Which other platforms fall in the same bucket?

The platforms that fall in the same bucket as Nutanix are the soft partitioning hypervisors Oracle does not recognise, including VMware vSphere, Microsoft Hyper V, and general KVM distributions. Oracle's accepted hard partitioning list is short and specific, covering technologies such as Oracle VM with CPUs pinned and capped to the policy rules and certain physical partitioning on some Solaris and IBM systems. Anything outside that list is, in Oracle's policy framing, soft partitioning, and the cluster wide exposure applies. For the accepted list in detail, read hard partitioning technologies Oracle accepts.

Indicative policy treatment. Confirm the current policy and your contract.
PlatformTypical policy treatment
Nutanix AHVNot recognised as hard partitioning
VMware vSphereNot recognised as hard partitioning
Microsoft Hyper V, general KVMNot recognised as hard partitioning
Oracle VM, CPUs pinned and cappedGenerally accepted as hard partitioning

For the parallel argument on the Microsoft and open source hypervisors, read Hyper V and KVM under Oracle's policy. The common thread across all of them is that the cluster wide claim depends on the policy, and the policy depends on your contract adopting it.

How do you design a Nutanix estate to limit exposure?

You limit exposure on Nutanix by isolating the Oracle workload to a defined set of nodes, controlling where it can move, and documenting those controls so the configuration is evidenced rather than asserted. A dedicated Oracle cluster, separated from the wider Nutanix estate, gives the cleanest position, because it removes the argument that the workload could move across a large pool of hosts. Where a dedicated cluster is not possible, restrict live migration, pin the workload, and keep records that show those restrictions were in place over the audited period. Evidence of control is what turns a contract argument into a defensible one.

None of this changes Oracle's policy, and you should not design around a hope that Oracle will quietly accept the platform as hard partitioning. The design exists to support the contract argument. If your agreement does not adopt the partitioning policy, a finding that licenses every node should narrow to the nodes where Oracle is installed and running, and a clean, documented configuration makes that narrowing credible. This is a contract dependent outcome, so the configuration and the agreement have to be read together.

Download the survival guide

Our VMware licensing survival guide covers Nutanix and the other unrecognised platforms, with a method to read a cluster wide claim against your contract. We reduce your Oracle exposure or we reimburse our service fee, on a Fixed Fee or Gainshare basis with no risk to you.

What is the buyer move?

The buyer move is to design Oracle on Nutanix for isolation and evidence, then meet any cluster wide claim by reading your contract before conceding a single node. Do not argue the platform. Argue that the policy driving the all node number is not a term of your agreement, and that the agreement licenses installed and running processors. If you are deciding where to run Oracle, weigh accepted hard partitioning against the documentation burden of the unrecognised platforms. For the full method, work up to the Oracle virtualization licensing guide, and if you want this assessed against your estate, our virtualization and cloud licensing service can scope it. To put a finding in front of us, use the contact form.

FAQ

How does Oracle treat Nutanix? Like VMware. The partitioning policy does not recognise Nutanix AHV as hard partitioning, which is the basis for cluster wide claims. The policy is not the contract.

Must I license the whole Nutanix cluster? Only if your contract adopts the policy. Otherwise the claim may narrow to the nodes where Oracle is installed and running.

Which platforms does Oracle accept? A short list including Oracle VM with pinned and capped CPUs and certain physical partitions, but not Nutanix, VMware, Hyper V, or general KVM.

Next step

Scope your Nutanix exposure before Oracle does.

Get a quote for a buyer side review of your virtualized Oracle estate and the cluster wide claim it might attract.

The License Position

Read Oracle's next move before they make it.

A short weekly note, buyer side. One development, why it matters, and one move you can make this week.

Buyer side only. Unsubscribe anytime.