Managed virtual server metrics.
Not every IBM product is priced by processor cores. A managed virtual server metric counts the virtual servers a product manages, and it behaves nothing like PVU. Conflate the two and you can over license by an order of magnitude, or hand IBM a finding by miscounting what is in scope.
IBM uses more than one way to count. PVU is capacity based and tied to processor cores. A managed virtual server metric is a usage based measure, a form of Resource Value Unit, that counts the virtual servers a product is managing or monitoring rather than the hardware that product runs on. The distinction matters because the same estate produces wildly different numbers depending on which metric the product actually uses.
What a managed virtual server metric counts
For management, monitoring, and automation tools, the value the software delivers scales with how many environments it looks after, not with the cores of the single server it runs from. So IBM counts the managed objects: the virtual servers under management. The licensing requirement grows as you point the tool at more of the estate, and it is largely indifferent to the size of the host the management server sits on. This is the opposite of PVU, where the host hardware is the whole story.
The Resource Value Unit family, including managed virtual server style metrics, is usage based and tied to a resource quantity such as managed servers, devices, or users, not to CPU. Pricing one of these products as if it were PVU, or the reverse, produces a number that is wrong on its face.
Where managed server counts go wrong
- Discovery scope mistaken for entitlement. A tool that can see a virtual server is not always managing it, but broad discovery can be read as a higher managed count than the entitlement covers.
- Decommissioned servers still counted. Virtual servers that were retired but never removed from the management inventory keep inflating the number.
- Tiering and role rules ignored. Some metrics distinguish how a resource is managed, and flattening those distinctions overstates the requirement.
- Wrong metric applied entirely. A product is reconciled against PVU when it is actually licensed by managed virtual server, or the reverse, so the entire calculation rests on the wrong unit.
How we reconcile a managed server metric
The work starts by confirming the metric each product is genuinely licensed under, because that decides everything downstream. For products on a managed virtual server measure, we separate what the tool can discover from what it is actually entitled to manage, strip out decommissioned and duplicate entries, and apply any tiering rules the metric defines. The defensible count is the set of virtual servers genuinely under management during the period, documented against the inventory, rather than the gross reach of the tool. Where IBM has applied PVU logic to an RVU product or the reverse, correcting the metric itself is often the single largest adjustment available.
Confirm the metric before you count. Managed virtual server and other RVU measures count managed resources, not cores, and PVU counts cores, not managed resources. Establish which unit governs each product, then count only what is genuinely in scope, and a number IBM presents as large can collapse to what the entitlement actually requires.