Zhivko Todorov
ALL CASE STUDIES

CASE 126 · SPARTAN · 2025

AURORAIO-OPTIMIZEDSTORAGECOST

Aurora I/O-Optimized for the workload, Standard for the rest.

A trading systems company ran a heavy Aurora Postgres cluster — high IOPS, predictable I/O cost. The team had stayed on Aurora Standard because they didn’t know what Aurora I/O-Optimized was. We modelled the workloads and migrated the right clusters to I/O-Optimized.

INDUSTRY

Trading systems

DOMAIN

COST

DELIVERED

2025

STACK

AURORA POSTGRES I/O-OPTIMIZED·AURORA STANDARD·CLOUDWATCH PERFORMANCE INSIGHTS·CUR + ATHENA

RESULTS

What changed, by the numbers.

AURORA COST

−27%

ON MIGRATED CLUSTERS

IO-OPTIMIZED ELIGIBLE

6 / 11

OF CLUSTERS

I/O BILL VOLATILITY

FLAT

NO MORE I/O SPIKES

MIGRATION DOWNTIME

0

AURORA CONFIGURATION CHANGE

HOW IT WENT

Aurora I/O-Optimized has a flat per-instance price including all I/O; Aurora Standard bills I/O separately and is cheaper for low-I/O workloads. The break-even depends on the workload’s I/O-to-instance-cost ratio. We modelled each cluster against the threshold using six months of CloudWatch metrics.

Six of eleven clusters came out clearly on the I/O-Optimized side — heavy I/O, predictable, the kind of workload Aurora Standard punishes you for. The migration was an in-place configuration change with no downtime. The other five clusters stayed on Standard because their I/O was modest.

On the migrated 6, Aurora cost dropped 27%. The bigger psychological win was flat I/O cost — no more month-end I/O bill spikes when trading volume increased. Finance stopped getting surprised by Aurora line items.

READY WHEN YOU ARE

Let's get your AWS bill (and architecture) in order.

The discovery call is free. You walk away with at least one concrete idea — even if we never work together.

Or email directly →