Executive Summary
A rapidly growing U.S.-based technology company providing enterprise software solutions faced an $82 million IBM licence compliance claim. The company operated a hybrid infrastructure spanning on-premises data centres, private cloud, and Kubernetes-orchestrated container environments. IBM products under audit included WebSphere Liberty (200+ instances, PVU metric), Cloud Pak for Integration (Kubernetes-native, VPC metric), DB2 (PVU metric), Tivoli (PVU/RVU metric), and Instana (per-host metric). The company had ILMT deployed on 70% of its virtual estate; the remaining 30% without ILMT agents triggered full-capacity pricing defaults. Within 3 months, Redress Compliance reduced the claim from $82 million to $600,000.
What Drove the $82M Claim
IBM's audit claim comprised three primary components, each reflecting systematic overstatement rather than genuine non-compliance. Understanding the drivers reveals why this claim was fundamentally contestable.
| Claim Component | Amount | Root Cause |
|---|---|---|
| Full-Capacity from ILMT Gaps | ~$58M | ILMT coverage at 70% triggered full-capacity multipliers (10-20x) on 30% of uncovered virtual machines |
| Kubernetes Node Licensing | ~$14M | Cloud Pak for Integration licensed at full node capacity rather than actual pod resource consumption |
| Legacy PVU/RVU Misalignment | ~$7M | Outdated processor values applied without crediting prior-period entitlements under legacy metrics |
Component 1: Full-Capacity Pricing from ILMT Gaps
IBM's sub-capacity licensing rules require 100% ILMT coverage. Any virtual machine without an ILMT agent defaults to full-capacity pricing at 10 to 20 times the sub-capacity rate. With 30% of the estate uncovered, IBM applied full-capacity multipliers across the entire uncovered segment, inflating the claim by approximately $58 million.
This was the largest single component because the multiplier effect compounds across hundreds of virtual machines. A single ILMT gap on a 16-core physical system becomes an assumed 64-core full-capacity allocation for licensing purposes. When extrapolated across the 30% uncovered portion of the estate, the impact was catastrophic.
Component 2: Kubernetes Container Licensing
IBM's audit applied full-node licensing to Kubernetes environments rather than pod-level resource allocation. This methodology treated every physical node in a Kubernetes cluster as fully licensed, regardless of actual container resource consumption. In practice, a 32-core node running three containers with 4-core resource requests was licensed at 32-core capacity.
The audit applied this full-node methodology to all Cloud Pak for Integration deployments, driving approximately $14 million in inflated exposure. This component was highly defensible because actual Kubernetes pod resource allocation data contradicted the auditor's node-level assumptions.
Component 3: Legacy PVU/RVU Metric Misalignment
IBM audit applied current PVU values to deployments without reconciling against historical purchasing under prior RVU or legacy PVU metric frameworks. The organization had accumulated significant entitlements under older metric frameworks that remained valid under the contract terms. These uncredited entitlements accounted for approximately $7 million in claim overstatement.
The Defence: Three Phases
Phase 1: ILMT Remediation and Sub-Capacity Validation
We deployed ILMT agents to the uncovered 30% of the estate and validated actual processor values for each IBM product. This critical step demonstrated that the actual sub-capacity position was a fraction of IBM's claim. Properly configured ILMT reporting showed that many virtual machines were consuming significantly fewer processor units than the full-capacity assumption.
This single phase eliminated the largest component of the claim. Once ILMT coverage reached 100%, the full-capacity defaults disappeared entirely. We presented IBM with the corrected ILMT data showing actual processor consumption across the entire estate, establishing a defensible sub-capacity baseline that IBM could not reasonably dispute.
Phase 2: Kubernetes Licensing Methodology Challenge
We presented IBM with evidence that pod-level resource allocation data, not node-level assumptions, was the correct basis for Cloud Pak for Integration licensing in Kubernetes environments. We provided detailed pod resource utilization reports showing actual VPC consumption against the over-stated node licensing position.
This challenge was grounded in IBM's own licensing documentation for Kubernetes-native products, which explicitly reference pod-level resource allocation. By showing that the auditor had applied an incorrect methodology, we forced a recalculation based on actual container resource consumption. The corrected calculation reduced this component from $14 million to a modest gap.
Phase 3: Legacy Entitlement Reconciliation
We reviewed all historic purchasing agreements, ELA renewals, and metric transitions. We reconciled current deployments against the full entitlement history, crediting PVU and RVU entitlements that IBM's audit had omitted. Every recovered entitlement reduced the effective licence gap. We documented that significant purchase history remained valid under the contract terms and should have been credited in the audit calculation.
Outcome Analysis
| Metric | IBM Claim | Final Outcome | Impact |
|---|---|---|---|
| Total exposure | $82,000,000 | $600,000 | 99.3% reduction |
| Full-capacity from ILMT gaps | ~$58M | Eliminated | Largest component removed |
| Kubernetes node licensing | ~$14M | Reduced to actual usage | Methodology corrected |
| Legacy PVU/RVU misalignment | ~$7M | Reconciled & credited | Entitlements recovered |
| Genuine licence gap | — | $600K | Modest, commercially resolved |
| Resolution timeline | — | 3 months | Rapid settlement |
Key Takeaways for Technology Companies
This engagement established several principles that apply across the technology industry, particularly for companies operating hybrid and Kubernetes environments:
- IBM audit claims against hybrid and Kubernetes environments are systematically overstated. Full-capacity assumptions are default audit methodology, not validated findings. Sub-capacity remediation is the fastest path to claim reduction.
- ILMT coverage gaps are the single largest driver of inflated IBM audit claims. A properly instrumented environment reduces exposure by orders of magnitude. If you operate with incomplete ILMT coverage, you are carrying unnecessary audit risk.
- Kubernetes licensing methodology is frequently applied incorrectly by IBM auditors. Pod-level resource allocation, not node-level assumptions, is the correct basis for licensing. Auditors often default to node-level methodology out of unfamiliarity with container orchestration metrics.
- Legacy entitlements are systematically uncredited in audit calculations. A systematic contract review uncovers millions in uncredited purchases from prior periods. If you have accumulated IBM contracts over multiple acquisition or refresh cycles, significant entitlements likely remain undocumented in current licensing records.
Why Speed Matters in IBM Audit Defence
The 3-month resolution timeline was not incidental. When an IBM audit claim is outstanding, it creates financial reporting uncertainty, consumes executive bandwidth, and creates strategic risk in vendor negotiations. Rapid, evidence-based defence is superior to prolonged dispute.
Our methodology achieves speed because it is grounded in technical remediation (ILMT deployment, methodology validation) rather than argumentation. When you present IBM with corrected ILMT data, actual pod resource allocation reports, and documented entitlement history, you create conditions for rational negotiation rather than protracted dispute.
The cost of defence was recovered multiple times over in the claim reduction. More importantly, the speed of resolution prevented months of financial uncertainty and enabled the organization to move forward with confidence in its IBM licensing position.