Zhivko Todorov
ALL CASE STUDIES

CASE 120 · DRIFTWOOD · 2023

NAT INSTANCENAT GATEWAYEGRESSCOST

NAT Gateway, replaced by an instance where the math says.

An internal IT team ran low-throughput VPCs (single-AZ test environments, internal dev clusters) where the NAT Gateway hourly cost dominated the real network usage. We replaced 14 of them with EC2-based NAT instances on t4g.nano with appropriate guardrails.

INDUSTRY

Internal IT

DOMAIN

COST

DELIVERED

2023

STACK

NAT INSTANCE·EC2 t4g.nano·AUTO SCALING·CLOUDWATCH METRICS·SSM

RESULTS

What changed, by the numbers.

PER-VPC NETWORK COST

−83%

TEST/DEV ENVIRONMENTS

VPCS MIGRATED

14

WHERE MATH FAVOURED IT

PRODUCTION VPCS

0 CHANGED

NAT GW STAYS

ANNUAL SAVING

$48K

NET OF ENGAGEMENT FEE

HOW IT WENT

NAT Gateway is the right answer for production — managed availability, automatic scaling, no maintenance burden. For low-volume internal environments, the hourly base cost was overwhelming the per-GB data-processing charge by 20x. NAT instances are old-school but still cheap.

We deployed t4g.nano NAT instances with Auto Scaling for replacement (not for capacity — these were single-instance setups). SSM handled patch management. CloudWatch metrics surfaced if any instance got close to saturation; the team would migrate that VPC back to NAT Gateway if so.

Per-VPC network cost on the migrated 14 environments dropped 83%. Production VPCs stayed on NAT Gateway untouched — we are not in the business of trading reliability for cost on the customer surface. The annual saving netted out at $48k after the engagement fee.

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 →