① Explainer · Licensing concepts
For most IBM mainframe estates the machine is far larger than the workload peak, so the choice between full capacity and sub-capacity licensing is often the single largest line on the bill. Here is the mechanism, a worked Rolling 4-Hour Average example, and the decision matrix for getting it right.
Two ways to be charged for the same software.
IBM Monthly License Charge software, the z/OS stack including CICS, Db2 for z/OS, IMS, and MQ, can be billed two ways. Full capacity charges on the total rated MSU of the machine the product can run on, whether or not you use it. Sub-capacity charges on the actual peak utilization of the logical partitions where the product runs, measured as a Rolling 4-Hour Average, or R4HA. Because the physical machine is almost always provisioned well above the everyday peak, sub-capacity is materially cheaper for most estates, and the difference between the two methods is frequently the biggest number a buyer can move.
Sub-capacity is not automatic. It is conditional on running IBM's free Sub-Capacity Reporting Tool, SCRT, and submitting the monthly report on time. The tool reads the SMF utilization data, calculates the R4HA for each eligible product, and produces the report that becomes the basis for the charge. Miss the report and you default to full capacity billing for that period. This explainer pairs with our MSU optimization service, where peak shaping turns the mechanism into savings.
The two methods compared on the dimensions that decide the bill
| Dimension | Full capacity | Sub-capacity |
|---|---|---|
| Billing basis | Total rated MSU of the machine | R4HA peak of the LPARs running the product |
| Reporting required | None; it is the default | Monthly SCRT report submitted on time |
| Typical cost | Higher; you pay for idle capacity | Lower; you pay for the measured peak |
| Example metrics | PSLC, zELC | WLC, AWLC, EWLC, and successors |
| Buyer levers | Few; the machine size sets the bill | Peak shaping, capping, zIIP offload, accurate reporting |
| Main risk | Overpaying for unused capacity | A missed report defaulting you to full capacity |
The same product, two very different bills.
Take a single machine rated at 1,000 MSU running a production z/OS LPAR. SCRT samples utilization every five minutes and rolls a four hour average across the month. Suppose the highest R4HA the product reaches all month is 420 MSU during the overnight batch window. Under sub-capacity, that 420 MSU peak is the billable figure. Under full capacity, the same product bills against the full 1,000 MSU rating of the machine, because the method ignores how much you actually used.
The figures are illustrative, but the shape is typical: when the machine is provisioned well above the workload peak, sub-capacity removes the idle capacity from the bill. The 58 percent gap here is the value of reporting correctly, and it recurs every month. A single missed SCRT submission would bill that month at the full 1,000 MSU.
The single most expensive avoidable mistake is a late or missing report, which defaults that month to full capacity. The reporting calendar is a financial control, not an administrative chore.
A missed report is a full capacity bill.
Because the charge follows a four hour peak, moving discretionary batch out of the peak window lowers the billable R4HA even when total work is unchanged. Timing is a lever in its own right.
When work runs matters as much as how much.
Defined capacity and group capacity limits hold the R4HA below a chosen ceiling. Soft capping trades a little headroom for a predictable, lower bill on the products that justify it.
A ceiling you choose beats a peak you discover.
Work that runs on zIIP specialty engines does not count toward the general purpose MSU peak that drives the charge. Eligible workloads moved to zIIP lower the billable R4HA directly.
Offloaded MSU is MSU you do not pay MLC on.
Full capacity bills MLC software on the total rated MSU of the machine. Sub-capacity bills on the R4HA peak of the LPARs that run the product. The machine is usually much larger than the peak, so sub-capacity is materially cheaper, and the gap is often the biggest line on the bill.
The system samples MSU every five minutes and averages continuously over a moving four hour window. The highest value in the month sets the billable MSU per product. SCRT reads the SMF data, calculates it, and produces the report you submit.
You default to full capacity for that period. Sub-capacity is conditional on running SCRT and submitting on time. A single lapsed month can be billed as if every eligible product ran at the full machine rating, which makes the calendar a financial control.
Shape the peak. Move discretionary batch out of the four hour window, cap LPARs with defined or group capacity, and offload eligible work to zIIP, which does not count toward the general purpose MSU peak. Peak shaping is the core of MSU optimization.
Related: licensing concepts hub · IBM licensing hub · IBM audit defense · MSU optimization
Audit notice or renewal under 18 months out? We mobilize within 48 hours.
Get expert help →