① Explainer · Sysplex Pricing
Combining machines in a qualifying Parallel Sysplex prices their MSUs as one larger figure, and because rates are regressive, that lowers the effective per MSU cost. But the 50 percent workload rule and full capacity metrics can quietly reverse the gain. This is a model, not a default.
Mainframe software pricing is regressive. The per MSU rate falls as the MSU total rises through successive bands, so the same workload is cheaper per unit at a higher aggregate level. Sysplex aggregation exploits this by letting you combine the eligible MSU capacity of multiple machines in a qualifying Parallel Sysplex and price the total as one larger figure rather than several smaller standalone figures. The combined total reaches deeper into the regressive schedule, and the effective rate across the estate drops.
Standalone is the opposite position. Each machine is priced on its own MSU figure, so two boxes of 300 MSU are each rated in the lower MSU bands at the higher per unit rate, instead of one 600 MSU figure rated partly in the cheaper bands. For multi machine estates, that difference compounds across every Monthly License Charge product on the floor.
Take two machines of 300 MSU each. Standalone, each is priced from the top of a regressive schedule on its own. Aggregated, the 600 MSU total runs further down the same schedule, so more of the capacity is priced in the cheaper bands. The table uses an illustrative regressive schedule to show the effect on the blended per MSU rate.
| Basis | MSU priced | Blended rate index | Relative cost |
|---|---|---|---|
| Machine A standalone | 300 | 100 | 300 units |
| Machine B standalone | 300 | 100 | 300 units |
| Standalone total | 600 | 100 | 600 units |
| Aggregated 600 MSU | 600 | 86 | 516 units |
Aggregating the same 600 MSU into one priced figure pulls the blended rate index from 100 to 86 in this illustration, roughly a 14 percent reduction on the combined MLC, because more of the capacity falls into the cheaper bands of the regressive schedule. The real number depends on your bands and metric. The mechanic, deeper bands at a lower rate, is the point.
| Dimension | Aggregated sysplex | Standalone |
|---|---|---|
| Per MSU rate | Lower, reaches deeper bands | Higher, each box priced alone |
| Capacity basis | Some metrics full capacity (PSLC) | Sub-capacity WLC available |
| Qualifying conditions | Coupling links, shared CF, workload share | None to maintain |
| Workload variability fit | Weaker if full capacity priced | Stronger under sub-capacity |
| Eligibility risk | Config change can break it | Stable, no threshold to hold |
| Best for | Steady, high MSU estates | Variable or smaller estates |
Two catches recur. First, qualification must hold continuously. A common requirement is that the sysplex LPARs running the systems enablement function represent at least 50 percent of the workload on each machine, alongside coupling links to a shared Coupling Facility. A configuration change can quietly drop you below the threshold and collapse the aggregation. Second, some aggregation metrics, notably Parallel Sysplex License Charges (PSLC), price at full capacity rather than the Rolling 4-Hour Average, so a lower headline rate can be offset by losing sub-capacity billing on variable workloads.
Buyer side levers
Model it. Do not default to it.
Related explainers: the MIPS to MSU conversion question, batch window tuning to cut R4HA peaks, and EWLC and entry workload license charges. Long tail product: CICS Transaction Server licensing. Commercial: cost optimization and IBM renewal advisory. Hub: the IBM buyer side guide.
Audit notice or renewal under 18 months out? We mobilize within 48 hours.