PVU ties cost to hardware capacity. RVU ties cost to something you use: managed devices, client endpoints, users, processors managed, or another resource defined for the specific product. The unit is the resource, not the CPU. That single difference changes where the audit risk lives. With PVU the risk is in virtualization and sub-capacity. With RVU the risk is in the count, because the count is a moving number that almost always drifts upward over the life of the entitlement.
What RVU actually measures.
An RVU entitlement licenses a defined quantity of a named resource. The exact resource depends on the product, but the pattern is consistent: you count the resource the product manages or serves, and you hold entitlement for at least that many units. Common bases include managed client devices, managed users, and managed processors, depending on what the product does. The product documentation defines the resource precisely, and that definition is the thing the auditor will hold you to.
How RVU differs from PVU.
The two metrics fail in different places, so they are defended differently:
- PVU is hardware bound. Cost follows cores and the per core value, and the main lever is sub-capacity. Virtualization rules decide the number.
- RVU is usage bound. Cost follows a resource count, and there is usually no sub-capacity equivalent to lean on. The lever is accurate, current counting and the tier you land in.
- RVU often uses tiered bands. The per unit treatment can step as volume rises, so the same overage can land in a different band than you expect.
Because RVU has no virtualization discount to reclaim, the defense is almost entirely about the integrity of the count and the entitlement records behind it.
The audit traps unique to RVU.
RVU exposure builds without anyone touching a server. The recurring traps:
- Quiet resource growth. Devices, users or managed endpoints grow over time while the entitlement was sized for day one.
- Counting the wrong thing. The product defines the resource precisely. Counting active users when the metric is managed users, or the reverse, misstates the position.
- Tier drift. Crossing a volume band changes how the resource is treated, and a count taken casually can sit in the wrong band.
- Stale baselines. An entitlement that matched the estate two years ago is reconciled against today's count at audit, not the count at purchase.
Reconciling an RVU position.
An RVU reconciliation starts with the product definition of the resource, then counts that resource as the auditor would, against current state. From there it maps the count to entitlements and Passport Advantage records, checks which tier the volume falls in, and identifies where growth has outpaced what was bought. Where the count is genuinely lower than the auditor's first reading, the work is to prove it with clean records before the higher number sets the baseline for the settlement.