Mainframe Cost Allocation

By Steven Thomas, Principal Evangelist

TL;DR

Mainframe CPU costs are often treated as a “black box.” By linking technical SMF data to business units and allocating costs fairly, organizations gain transparency, encourage cost-saving behavior, and can save millions annually

• The challenge: Mainframe CPU costs are massive (often tens or hundreds of millions per year) but poorly understood and rarely allocated correctly.

• Business mapping: Translating SMF metrics (job names, transactions, report classes) into business dimensions like applications or units creates transparency.

• Handling peaks: CPU costs are driven by peak utilization; cost models must reflect both fairness and stability across peak and off-peak usage.

• Fair allocation: Explicit mapping plus proportional distribution of overhead ensures costs are both accurate and accepted by stakeholders.

• Self-service reporting: Enabling stakeholders to drill down into their own costs fosters accountability and reduces disputes.

• Implementation: Cost allocation is a project requiring executive sponsorship, strong reporting tools, and clear processes for show-back and eventual charge-back.

About

Mainframe customers often spend tens or even hundreds of millions of euros each year on CPU-related costs. For in-house environments, these costs come from hardware and IBM software charges; for outsourced customers, service providers typically bill based on CPU usage.

Too often, these expenses are treated as a black box. Cost allocation to applications or business units—if it happens at all—is usually rudimentary. This leaves the people driving the costs with little transparency, no clear way to reduce them, and limited motivation to act. Even when they try, efforts often miss the real cost drivers and fail to deliver savings.

Done correctly, cost allocation changes this. By linking usage to business activity, organizations gain visibility, align costs with responsibility, and unlock massive savings.

Read more in the whitepaper.