① Journal · IBM · COBOL Compiler Family
The COBOL compiler family is an IPLA one time charge with an annual support stream, so the renewal number moves on entitlement, sub-capacity tier, and version, not the list rate. Here are the seven levers that actually shift a compiler renewal, and how each one works.
The list rate is the negotiation people expect. The entitlement and the support stream are the ones that decide the bill.
The COBOL compiler family, IBM Enterprise COBOL and its companion compilers such as Enterprise PL/I, is licensed under IBM's International Program License Agreement as a one time charge plus an annual Subscription and Support stream, with the entitlement sized to the MSU sub-capacity tier of the LPARs where the compiler is licensed. The recurring number on a renewal is the support, and it is driven by how much capacity you are entitled on and how many versions and product entitlements remain under support, not by a per developer rate. Buyers expect to argue the list, but the levers that move a compiler renewal right size the entitlement and prune what is still carried.
The structural fact behind every lever is that the runtime is not the compiler. The COBOL runtime ships in z/OS through Language Environment, so compiled programs execute without a separate compiler license; you pay to build and recompile, not to run. That separates compile time capacity from the production estate and reframes the whole renewal. Read this with our explainer on MIPS and MSU, the Enterprise COBOL licensing page, and the IBM publisher hub.
What we commonly observe · the lever and how it shifts the compiler number
| Lever | How it works | What it moves |
|---|---|---|
| Separate compile time from run time | Runtime ships in z/OS, so license the build LPARs only | The capacity you must license at all |
| Right size the sub-capacity tier | Match entitled MSU to the development and build estate | The one time charge and the support base |
| Prune the version count | Drop old compiler entitlements no longer compiled against | The number of products under support |
| Test the case for support | Perpetual license keeps the last version without S&S | Whether the support stream is worth keeping |
| Model Value Unit Edition | Compare the VUE variant against standard IPLA where it fits | The cost basis for the entitlement |
| Time it to a hardware event | Align the renewal with a processor refresh or contract reset | The leverage you hold at the table |
| Validate entitlement vs deployment | Reconcile the sub-capacity report against where it runs | The accuracy of the number everything rests on |
Levers are what we commonly observe working across buyer side engagements, not statements of IBM policy, and the effect of each depends on your build cadence and hardware roadmap. Your entitlement, deployment, and version mix govern the real number.
The COBOL runtime ships in z/OS through Language Environment, so compiled programs run without a compiler license. You pay to build and recompile, which means the entitlement only needs to cover the development and build LPARs, not the production systems that merely execute the output. Confining the licensed capacity to where compilation actually happens is the single largest structural lever on a compiler renewal.
You license the act of building, not the act of running.
Estates accumulate compiler entitlements: old Enterprise COBOL releases, retired PL/I products, capacity that no longer matches the build footprint. Each one carries support. Reconcile the entitlement against where compilation actually happens, drop the versions you no longer compile against, and align the entitled MSU to the real development estate before the support base is recalculated.
Support you renew on shelfware is the easiest line to cut.
An IPLA license is perpetual, so dropping Subscription and Support keeps the right to run the last entitled version while forgoing fixes, new releases, and exploitation of future processors. For a stable estate not tracking the newest language level the trade can favor reducing support; for one chasing the latest hardware it usually does not. Decide on your build cadence and hardware roadmap, modeled against the support you would give up.
Keep support where it buys something, drop it where it does not.
④ Where the compiler number is moved
The list rate is the argument you expect. The entitlement, the versions, the support are the ones that count. License the build, prune the versions, test the support.
Typical reduction negotiated on renewal spend
Mainframe spend negotiated on the buyer side
Engagements delivered since 2019
The compilers are IPLA products: a one time charge plus annual Subscription and Support sized to the MSU sub-capacity tier of the LPARs they are licensed on. The recurring number is the support, driven by entitled capacity and the count of versions and entitlements still carried, not a per developer rate. See Enterprise COBOL licensing.
Not for the runtime. The COBOL runtime ships in z/OS through Language Environment, so compiled programs execute without a compiler license. You license the compiler to build and recompile, which confines the entitlement to the development and build LPARs. See MIPS and MSU explained.
Often, with care. The perpetual license keeps the right to run the last entitled version, but dropping support forgoes fixes, new releases, and future hardware exploitation. For a stable estate the trade can favor reducing support; for one tracking the newest hardware it usually does not. Model it against your build cadence. See reinstatement fees after lapsed support.
Confine the entitlement to the build LPARs, prune retired versions, test the case for support, model Value Unit Edition where it fits, time it to a hardware event, and validate entitlement against deployment. Our license negotiation service sets the structure; see also IBM price increase patterns.
Related: IBM publisher hub · Enterprise COBOL licensing · Db2 for z/OS renewal negotiation · IBM price increase patterns · license negotiation
Audit notice or renewal under 18 months out? We mobilize within 48 hours.
Get expert help →