The Challenge

A major Australian retail bank, ranked in the top five by asset value and operating across 15,000+ employees nationwide, maintained one of the country's largest and most complex IBM estates. The infrastructure spanned IBM z/OS mainframe systems, IBM Db2 database platforms, CICS transaction servers, IBM MQ message queues, and a distributed middleware ecosystem across virtualised server environments.

In late 2025, IBM initiated a formal Software Asset Management (SAM) review—the precursor to a full compliance audit. IBM's preliminary data request demanded three years of complete deployment history, configuration records, and licence usage metrics across the entire estate. The scope was comprehensive: every mainframe LPAR, every distributed server, every database instance, and every middleware component.

The bank's immediate problem was visibility. While mainframe licensing was relatively well-documented, the distributed server infrastructure had significant gaps in IBM Licence Metric Tool (ILMT) coverage. ILMT agents were either absent, misconfigured, or failing to report usage across 94 virtualised servers hosting IBM middleware, Db2, and integration platforms. Without accurate, auditable deployment and usage metrics, the bank faced IBM's worst-case measurement scenario: IBM would extrapolate from partial data or apply conservative assumptions that inflated licence requirements.

IBM's preliminary claim reached AUD $8.3M (approximately USD $5.5M), driven by aggressive PVU (Processor Value Unit) calculations on virtualised infrastructure, assertions about processors that had never been deployed, and overstated concurrent user counts on middleware platforms.

"Audit defence is not about disputing IBM's authority to conduct audits—it's about ensuring they measure accurately. When ILMT data is missing, vendors fill the gap with assumptions that favour their licensing position. We deployed the right tools and processes retroactively, backed every claim with evidence, and negotiated from a position of demonstrated compliance."

The Approach

Redress was engaged to mount a comprehensive audit defence strategy across three concurrent work streams: immediate ILMT deployment, rigorous measurement methodology challenge, and structured settlement negotiation.

ILMT Deployment at Scale (Weeks 1-7): The bank's IT teams had never fully deployed ILMT across the distributed estate. We initiated a coordinated rollout: ILMT agents were installed on all 94 virtualised servers, configured for continuous reporting, and integrated with IBM's reconciliation systems. Within seven weeks, we had complete, auditable deployment and usage metrics covering a 24-month lookback period. This retroactive data proved critical: it documented actual processor allocations, real user concurrency metrics, and precise middleware deployment footprints. The evidence contradicted IBM's assumptions at almost every level.

Measurement Methodology Challenge: IBM's PVU calculation method contained several errors. First, they claimed processors that were never deployed in the customer's infrastructure. Second, they applied full licence requirements to virtualised environments where only a fraction of processors were actually allocated to IBM software at any given time. Third, they used maximum observed user counts rather than average concurrent usage for middleware platforms. We documented each error with forensic precision: server configuration reports, hypervisor allocation records, and ILMT metrics showed that IBM's claim overstated actual usage by a factor of three.

Negotiation and Settlement: Armed with complete ILMT data and a detailed methodology challenge, we initiated settlement discussions with IBM's licensing executive team. The conversation shifted from "you owe $8.3M" to "here's what you actually consumed, measured to the day." IBM's position weakened immediately: they had no counter-evidence to the ILMT data. Over four weeks of structured negotiation, we worked through IBM's cost allocation model, challenged their PVU tables, and demonstrated the bank's genuine commitment to compliance. The outcome: IBM withdrew the $8.3M exposure, the bank's existing ELA was restructured with clear compliance safe harbours and automated reporting obligations, and no additional licence purchase was required.

The Outcome

The case closed in March 2026 with a zero-cost settlement. The bank eliminated the entire AUD $8.3M exposure without purchasing additional licences or entering into emergency true-ups. More importantly, the organisation achieved permanent structural compliance: ILMT was fully operational across all 94 distributed servers, the ELA was reframed with explicit safe harbour clauses tied to ILMT reporting, and the customer's licensing risk profile shifted from high-exposure to fully-managed.

The PVU reduction of 67% reflects the gap between IBM's initial assumptions and actual measured usage. This metric is crucial because it demonstrates how easily SAM claims can become inflated when measurement data is incomplete. The seven-week timeline for ILMT deployment across 94 servers, while aggressive, was achievable because the bank's infrastructure was modern and cloud-native: agents deployed cleanly, virtualisation platforms supported agent reporting, and historical logs were retained for retrospective analysis.

The financial impact extended beyond the avoided $8.3M claim. The restructured ELA reduced annual maintenance costs by 12% through more efficient licence allocation and eliminated the bank's need for reactive compliance spending. More significantly, the bank gained a replicable audit defence playbook: future IBM reviews would trigger the same ILMT-first protocol, removing uncertainty from the audit relationship.

Key Takeaways

Measurement Gaps Are Your Greatest Risk: When ILMT data is absent or incomplete, vendors default to extrapolation and assumption. These assumptions almost always favour vendor revenue. Retroactive ILMT deployment, while challenging, turns the audit from a negotiation into a data-driven conversation.

PVU Calculations Are Contestable: Processor allocation, virtualisation multipliers, and licence stacking rules are not matters of vendor interpretation—they're defined in the licence agreement and documented in your infrastructure. Challenge every assumption with architectural evidence.

Scale Matters for Timeline: Deploying ILMT across 94 servers in seven weeks required focused coordination and dedicated resource. For larger estates, allow 10-14 weeks. For smaller configurations, the timeline compresses to 2-3 weeks. Plan accordingly, but do not delay—immediate ILMT deployment shifts the audit conversation in your favour.

Settlement Leverage Comes From Data: IBM's audit position is strongest when you lack evidence. The moment you produce complete, auditable ILMT data, their cost position weakens. Leverage that moment to restructure your ELA with compliance safe harbours and reporting obligations that give you ongoing visibility into your licence exposure.

Anonymisation Protects Your Negotiating Room: We've worked this case with anonymity to protect the bank's ongoing relationship with IBM. This approach is standard in audit defence: vendors are more willing to move toward commercial solutions when the conversation happens off the record. Ensure your legal and procurement teams understand that audit defence often requires discretion.

Facing an IBM SAM audit? Download our Audit Defence Kit for z/OS, Db2, and distributed middleware environments.

Complete measurement protocols, challenge frameworks, and negotiation templates.
Download Kit →