IBM software in Azure: licensing and ILMT coverage.

The metric does not change in the cloud

IBM software carries its licensing metric with it. A processor value unit product is still counted on the cores allocated to it, whether those cores sit in your data center or in an Azure virtual machine. Public cloud does not introduce a separate rulebook. What it changes is the practical question of how many cores the software is licensed against, and that turns entirely on whether your sub-capacity reporting reaches the cloud instances.

Sub-capacity licensing lets you license only the virtual cores running IBM software rather than the whole physical host. In Azure terms, that means the virtual cores assigned to the VM running the product, not the underlying Microsoft hardware you never see. But sub-capacity is conditional. It requires ILMT, or an approved tool such as Flexera One ITAM or HCL BigFix Inventory, deployed within 90 days of first eligible deployment, running continuously, with quarterly reports retained for two years. Those conditions do not pause because the workload is now in the cloud.

Where Azure deployments lose sub-capacity

• ILMT never followed the workload. The agent was running in the data center but the migrated Azure VMs were never added to inventory, so the cloud instances have no sub-capacity evidence at all.
• Autoscaled instances slip the count. VMs that spin up and down can run IBM software during a window ILMT did not capture, leaving gaps a discovery scan will surface.
• The 90 day clock was missed on the new instances. A fresh eligible deployment in Azure starts its own clock, and a tool that is not measuring it in time forfeits sub-capacity for that period.
• Reporting lapsed during the migration. Project pressure interrupts the quarterly report cadence, and the lookback later runs at full-capacity rates for every quarter the evidence is missing.

What auditors check for cloud workloads

An IBM audit treats cloud instances the same way it treats on-premise hosts: it asks for continuous ILMT data proving how many cores the software actually used. Where that data is missing or broken for a period, the claim defaults to full capacity for the affected instances, and the lookback can run two to five years of back-payments. The exposure on a busy cloud estate compounds quickly because the per-period gap repeats across every quarter the evidence does not cover. The defense is the same as on-premise too: prove continuous, correct measurement of the virtual cores, and the count drops back to what the software actually consumed.