The instinct is understandable. PVU exposure is frightening, audits are unpredictable, and a buffer of unused entitlement looks like insurance against an unexpected finding. The trouble is that this particular insurance does not pay out the way buyers expect. An audit does not net your total purchased PVU against your total deployed PVU. It checks each product against its own entitlement, and a surplus in one place does almost nothing for a shortfall in another.
The buffer that protects nothing.
When IBM reconciles an estate, it builds the finding from specific products on specific hosts measured against the entitlement for those exact products. A general reserve of PVU does not slot into that calculation. If the finding is on WebSphere and the surplus was bought against a different product, the surplus is invisible to the finding. The buffer was sized against a fear, not against the way the audit actually counts, so it sits unused while the gap it was supposed to cover is charged anyway.
Why surplus does not offset a shortfall.
Entitlements are not fungible across products or, in many cases, across metrics. The reasons a buffer fails to transfer:
- Product specificity. Entitlement for one product does not satisfy a deployment of another. Extra Db2 does not cover a WebSphere gap.
- Metric mismatch. PVU surplus does not offset an RVU or user based shortfall. The units are not interchangeable.
- Bundle boundaries. Entitlement that came bundled with another product is constrained to its allowed scope, not a free pool to draw on.
- Version and edition limits. Entitlement for one edition does not automatically cover a higher edition that was actually deployed.
The legitimate offsets in an audit come from entitlements IBM overlooked for the very products in scope, not from unrelated surplus bought as a hedge.
The cost of carrying unused entitlements.
A buffer is not free to hold. Beyond the purchase, entitlements typically carry ongoing support and subscription charges that recur whether or not the capacity is ever used. So the cushion bought once keeps billing year after year, quietly, for protection it does not provide. Budget that could have funded accurate measurement or sub-capacity tooling instead funds shelfware that an auditor will never credit back to you.
The hedge that actually works.
The real protection against an unexpected finding is not surplus, it is knowing your position before IBM does. That means accurate, current measurement of what is deployed against what is entitled, product by product; a clean sub-capacity posture where PVU products are virtualized; and the discipline to reconcile on a schedule rather than discovering the gap in an audit letter. Money spent there reduces real exposure. Money spent on a just in case buffer mostly reduces the bank balance.