Mainframe MSU reduction, the largest saving and the least pursued
IBM mainframe software licensing runs on two related units: MSU, the software metric billed on the rolling four hour average peak, and MIPS, the hardware capacity view. Across our reviews, MSU and MIPS reduction was the largest single saving available and the least pursued, because the levers are operational before they are contractual, and the operational ones compound.
Prepared by Redress Compliance · August 7, 2026 · IBM advisory. Based on 20 to 30 IBM Z mainframe cost reviews worked 2024 to 2025.
Executive summary
The peak is the bill, and nobody was managing the peak. Sub capacity pricing bills the rolling four hour average peak per LPAR rather than machine capacity, cutting licensed MSU 30 to 60 percent for estates running multiple LPARs at different peak times, yet soft capping and Tailored Fit Pricing sat unconfigured in our reviews, leaving 10 to 25 percent of MSU spend unmanaged, and workload was rarely shifted off peak despite cuts of 8 to 18 percent sitting in the schedule.
The ILMT trap is measured in days and priced in millions. Sub capacity eligibility requires ILMT deployed and reporting continuously with monthly SCRT submissions by the ninth day and two years of retained reports: a single ILMT outage longer than seven days can flip the affected month to full capacity pricing, seven figures on a mid size estate, and most audit findings traced to ILMT gaps rather than under licensed software. The discipline is daily health checks, the incident playbook for any outage past twenty four hours, and the logged uptime that is the audit defense.
The operational levers run first because they compound. A workload shift dropping the R4HA peak ten percent reduces the bill for every product priced on it, so the sequence is fixed: workload shift at 5 to 15 percent, zIIP and zAAP offload at 10 to 25, subsystem decommissioning at 3 to 8, and defined capping at 5 to 10, with the contractual levers, Tailored Fit Pricing, the ELA negotiation at 10 to 25 percent on price, and bundling, then applying against a lower baseline, multiplying the saving.
The ELA renewal anchors the next three to five years. The MSU number at renewal is the baseline the whole term prices from, and the worked case shows the compounding: a renewal closed 31 percent below IBM's opening proposal, fourteen points from a twelve month ILMT discipline reset and workload shift program on the operational side, seventeen more from the contractual levers applied to the reduced base, timed to IBM's quarter end.
The reduction lever matrix, sequenced
| Lever | Category | Typical reduction | Effort |
|---|---|---|---|
| Workload shift to off peak | Operational | 5 to 15 percent of MSU | Medium |
| zIIP and zAAP offload | Operational | 10 to 25 percent of MSU | High |
| Decommission unused subsystems | Operational | 3 to 8 percent of MSU | Medium |
| Capping with defined limits | Operational | 5 to 10 percent of MSU | Medium |
| Tailored Fit Pricing | Contractual | 0 to 10 percent on price | Low |
| The ELA renewal negotiation | Contractual | 10 to 25 percent on price | High |
| Multi product bundling | Contractual | 5 to 15 percent on price | Medium |
The ILMT discipline, five practices and a playbook
- The daily health check: agent reporting confirmed on every monitored LPAR, because the seven day outage cliff is measured from the gap, not the discovery.
- The monthly SCRT submission: by the ninth day of the following month, no exceptions, feeding the utilization data sub capacity billing depends on.
- The quarterly reconciliation: ILMT inventory against Passport Advantage contract data, catching the drift before IBM does.
- The annual data quality review: independent third party validation of the ILMT record the audit will read.
- The incident playbook: a written response for any outage past twenty four hours, with the logged uptime as the audit trail that holds.
The IBM Z mainframe negotiation brief
The MSU mechanics, the pricing models, the ILMT discipline, and the seven ELA moves worked on a representative estate.
Get the white paper →The seven ELA moves, in renewal order
The ELA renewal is the largest single negotiation moment for any mainframe customer, and the sequence runs: baseline the MSU consumption from two years of SCRT data, peak by product and LPAR; run the operational lever audit, workload shifts and zIIP candidates, because the reduction belongs to you before it discounts for IBM; score the bundle, which products carry weight and which are paid but unused; pull the independent MSU rate benchmarks by product family; open the alternative, distributed migration scenarios for the workloads that could credibly move; cap the annual uplift at zero to three percent, fixed in the contract; and time the close to IBM's quarter ends, March, June, September, and December. The pricing model landscape underneath, WLC and VWLC for most z/OS estates through Tailored Fit Pricing for predictable growth, decides which meter the moves negotiate against, and the wider portfolio context sits in the ELA analysis and the mainframe CIO advisory.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across mainframe reviews, 2024 to 2025
Across roughly 20 to 30 IBM Z mainframe cost reviews Fredrik Filipsson worked between 2024 and 2025, MSU and MIPS reduction was the largest single saving available and the least pursued:
Soft capping and Tailored Fit Pricing left unconfigured on estates that qualified for both.
Operational points from ILMT discipline and workload shift, contractual points on the reduced base.
The Mizuho scale reference frames what the levers reach: the $71 million mainframe optimization case ran the same sequence at financial institution scale, and the sub capacity reporting gaps we kept finding, parts of estates defaulting to full machine capacity because the reporting chain broke, were the same failure at every size. The mainframe estate rewards exactly one thing: the operational discipline that manages the peak daily, feeding the contractual negotiation that anchors the baseline for the next three to five years.
Your first five moves
- Confirm ILMT health daily and log it, because the seven day outage flips a month to full capacity and the audit trail is the defense.
- Baseline the R4HA peaks from two years of SCRT data, peak by product and by LPAR, the number every lever works against.
- Shift workload off the peak window first, the 8 to 18 percent that cuts every product priced on the average at once.
- Configure the capping and evaluate Tailored Fit Pricing, the 10 to 25 percent of MSU spend running unmanaged.
- Run the seven ELA moves against the reduced baseline, capped at zero to three percent and timed to IBM's quarter. The IBM practice runs the estate with you.
Frequently asked questions
What is the difference between MSU and MIPS?
MSU, Million Service Units per hour, is IBM's software licensing metric, billed on the rolling four hour average peak under sub capacity pricing; MIPS, Million Instructions Per Second, is the hardware capacity view used for sizing. They move together but are not identical, and the software bill follows the MSU peak, which is why managing the peak is the whole game.
How does IBM mainframe sub capacity pricing work?
You pay for the peak rolling four hour average workload per LPAR rather than total machine capacity, cutting licensed MSU 30 to 60 percent for estates running multiple LPARs with staggered peaks. Eligibility requires ILMT deployed and reporting continuously, monthly SCRT submissions within nine days of month end, and two years of retained reports for audit verification.
What happens if ILMT goes down?
A single outage longer than seven days can flip the affected month to full capacity pricing, a seven figure exposure on a mid size estate, and most mainframe audit findings trace to ILMT gaps rather than under licensed software. The discipline is daily health checks, a written incident playbook for outages past twenty four hours, and logged uptime as the audit trail.
How do you reduce mainframe MSU costs?
Operational levers first, because the R4HA peak multiplies across every product priced on it: workload shifted off peak for 5 to 15 percent, zIIP and zAAP offload for 10 to 25, subsystem decommissioning and defined capping for the rest. The contractual levers, Tailored Fit Pricing and the ELA negotiation at 10 to 25 percent on price, then apply against the lower baseline.
What should an IBM mainframe ELA renewal achieve?
The worked case closed 31 percent below IBM's opening proposal: fourteen points from a twelve month ILMT discipline reset and workload shift program, seventeen more from the contractual levers on the reduced base, with the annual uplift capped at zero to three percent and the close timed to IBM's quarter end. The MSU number at renewal anchors the next three to five years.
Why is mainframe MSU reduction so rarely pursued?
Because the levers are operational before they are contractual, and they sit between teams: the schedulers who could shift workload, the engineers who could offload to zIIP, and the procurement team negotiating the renewal rarely share a baseline. In our reviews it was the largest single saving available on the estate precisely because nobody owned the peak.
