MIPS and workload metrics in an Enterprise Server audit

When a COBOL estate moves off the mainframe onto Enterprise Server, the licensing conversation often inherits mainframe units like MIPS, and that inheritance is where a finding can go badly wrong. MIPS and workload metrics in an Enterprise Server audit need careful handling, because a capacity figure carried over from a mainframe sizing exercise rarely describes the workload the distributed deployment actually runs.

Enterprise Server reached the OpenText estate through the Micro Focus acquisition that closed on January 31, 2023, and is governed by the Additional License Authorizations rather than the OpenText EULA. Enterprise Server lets COBOL and mainframe workloads run on distributed platforms, and the metrics that travel with that migration, MIPS equivalents, workload units, or capacity measures, are easy to misapply. A finding that prices the provisioned capacity of the host rather than the workload the application consumes will overstate the position, sometimes substantially.

What MIPS and workload metrics try to capture

MIPS, millions of instructions per second, is a mainframe capacity unit, and on distributed Enterprise Server deployments it or an equivalent workload measure is sometimes used to size the license. The metric tries to capture the processing the COBOL workload consumes. The trap is the difference between provisioned capacity and consumed workload: a host can be sized generously for peak or for future growth, and a finding that prices the provisioned capacity treats headroom the application never uses as licensed consumption. The authorization is the document that defines the workload unit and how it is measured, and that definition is the boundary the finding must respect.

This is the workload analogue of the core counting problem covered in how core based metrics inflate a Visual COBOL finding: in both cases a capacity figure stands in for actual use and inflates the number.

Where a workload metric inflates the finding

A handful of patterns push a MIPS or workload based Enterprise Server finding above the defensible figure. Each is a point to test against the agreement and the measured workload.

• Provisioned capacity priced as consumed. Charging the full capacity of the host rather than the workload the application ran over a representative period.
• Peak treated as steady state. Using a one time batch peak or a month end run as the baseline rather than the sustained workload.
• Non production workloads counted in. Including test, development, and disaster recovery capacity as though it were production, a subject we develop in COBOL non production and test environment scope.
• Mainframe sizing carried over uncritically. Importing a MIPS figure from the original mainframe contract without remeasuring on the distributed platform.

Reconstruct the workload against the authorization

The four Rs apply directly. Respond inside the seven day notice window and route everything through a single controlled channel, so the workload is described once rather than reconstructed by the vendor from a capacity scan. Reconstruct the effective position by reading the authorization for how the workload unit is defined and measured, then producing a workload record over a representative period to show actual consumption. Rebut the finding line by line where it prices provisioned capacity, treats a peak as steady state, or counts non production workloads. Resolve on terms that fix the workload measurement basis so the next audit does not reopen the same capacity. Building the workload evidence before any vendor measurement runs is decisive, and our note on reducing a COBOL finding with workload evidence covers how to assemble it.

In a recent engagement

In a recent Enterprise Server engagement, a finding rested on a MIPS equivalent figure carried over from the legacy mainframe contract and applied to a distributed host provisioned well above the workload it carried. Remeasuring the actual workload over a representative period, reading the authorization for how the metric was defined, and excluding the disaster recovery capacity that had been swept in brought the figure down to the workload the application genuinely consumed. The reduction came from the gap between provisioned capacity and measured workload, which is the gap most workload based findings rely on.

Hold the finding to the measured workload

With Enterprise Server more than most products, a MIPS or workload based finding depends on whether anyone insists that the metric prices consumed workload rather than provisioned capacity. A buyer that accepts an inherited mainframe sizing is paying for processing the distributed deployment never performs. The defensive discipline is to read the authorization for how the workload unit is defined, measure the actual workload over a representative period, and rebut every line that prices capacity or peaks as steady consumption. Most of the reduction available on a workload based Enterprise Server finding comes from establishing the measured workload and refusing to be charged for headroom.

Related field notes

These notes from the COBOL and Enterprise Server mainframe audit defense cluster cover workload, capacity, and MIPS metrics. Each links back to the complete OpenText audit defense playbook for 2026.

• Enterprise Server transaction and MIPS models
• how core based metrics inflate a Visual COBOL finding
• reducing a COBOL finding with workload evidence
• what is an Enterprise Server core license metric
• COBOL non production and test environment scope

If an OpenText or Micro Focus audit notice has arrived, the first seven days carry more weight than any week that follows. OpenText Audit Defense is an independent, buyer side practice founded in 2020 by former vendor compliance leadership. We have defended more than 200 audits, reduced the average finding by 68 percent, and mitigated more than $90M in claims against vendor positions. We do not resell OpenText software and we are not affiliated with OpenText Corporation. To open a case, use the contact form on this site.