Mainframe CPU Optimization for a US Life Insurer

Overview

A US-based life insurer's rapid growth was driving up mainframe operation, maintenance, and scaling costs. Complex calculations were run on every transaction, regardless of whether the underlying data changed, consuming CPU capacity without adding value.

More

A US-based life insurer's rapid growth was driving up mainframe operation, maintenance, and scaling costs. Complex calculations were run on every transaction, regardless of whether the underlying data changed, consuming CPU capacity without adding value.

Instead of a full infrastructure overhaul, Tech Mahindra deployed two targeted solutions that delivered an 8% enterprise-wide reduction in CPU usage, recovered capacity, stabilized the cost base, and created headroom for future growth.

Less
us-insurer-thumb

Industry Challenge

The insurer's operations, spanning investments, deferrals, pensions, and contract-to-plan processing, ran on a single mainframe infrastructure that processed millions of transactions daily. Rapid scaling created three pressures: rising infrastructure and expansion costs, redundant data-intensive calculations executed on every transaction across all five workloads, and unnecessary application calls during deferral retrievals consuming CPU beyond the required workload.

Two Targeted Fixes that Recovered Capacity

Tech Mahindra's approach centered on eliminating processing waste at the source rather than scaling infrastructure. Two targeted solutions, deployed simultaneously, reduced CPU consumption across the enterprise.

Answers at Rest

Computed outputs are stored in a dedicated Db2 performance table and reused rather than recalculated for every transaction. A publish/subscribe system refreshes stored results only when data changes. This fix is rolled out in four phases over 18 months.

Deferral Block Optimization

Analysis of deferral workloads during fetches revealed unnecessary application calls adding CPU strain. Eliminating these calls delivered a targeted reduction in program-level CPU utilization within the deferral block.

Benefits

  • 8% enterprise-wide CPU load reduction across the Sysplex
  • 64% decline in program-level CPU within the deferral workload
  • Maintained a stable cost base without new infrastructure investment
  • Maintained production stability throughout the phased 18-month rollout