DXC Technology — American Airlines Sabre PSS

## Context

Before Hack Reactor and Airbnb, I operated IBM Z-series mainframes running TPF (Transaction Processing Facility) for American Airlines' Sabre Passenger Service System — the reservation backbone processing millions of PNR (Passenger Name Record) transactions a day at sub-second latency.

## Approach

### Production monitoring

I drove production health monitoring across multiple LPARs with MainView, OmegaMon, and NetView — tracking CPU utilization, DASD performance, CICS region availability, and TPF transaction queue depth. Anomalies got escalated before they reached customer-facing booking flows.

### Batch and incident operations

I executed and monitored batch workloads via JCL and enterprise scheduling — overnight reservation reconciliation, end-of-day settlement, financial postings. During incidents and scheduled maintenance windows, I issued TPF console commands and navigated operator screens through 3270 terminal emulation (TN3270 / IBM Personal Communications), following runbook procedures under time pressure.

I coordinated ITIL-aligned incident and change management in ServiceNow, partnering with TPF assembler / SHARES developers and routing issues with diagnostic context — transaction traces, memory dumps — when a fix needed to go past the operator layer.

## Outcome

Tightened runbook ownership and automated repetitive operator workflows, reducing mean time to recovery on critical batch jobs. The role built direct, hands-on familiarity with TPF's constraints — single-process architecture, in-memory database, fixed-record blocking, assembler-era performance budgets — the kind of mission-critical legacy infrastructure most modern engineers never touch, and a different discipline of reliability than anything in the web stack.