Zhivko Todorov
ALL CASE STUDIES

CASE 119 · ANCHOR · 2025

CLOUDFRONTORIGIN SHIELDCACHEEGRESS

Origin Shield, sized to actually pay for itself.

A software distribution company served 14PB/month from CloudFront with origins in S3 across two regions. Cache hit rate at the edge was good (84%); origin fetches were still expensive because they hit S3 from every edge location independently. We added Origin Shield and brought origin fetches down 73%.

INDUSTRY

Software distribution

DOMAIN

COST

DELIVERED

2025

STACK

CLOUDFRONT·CLOUDFRONT ORIGIN SHIELD·S3·CLOUDFRONT FUNCTIONS

RESULTS

What changed, by the numbers.

ORIGIN FETCHES

−73%

S3 + EGRESS COMBINED

EFFECTIVE CACHE HIT RATE

96.4%

WAS 84%

S3 + EGRESS BILL

−$22K/mo

NET OF SHIELD COST

p50 LATENCY

−4%

WARMER MID-TIER

HOW IT WENT

CloudFront has roughly 600 edge locations. Without Origin Shield, a cache miss at any of them goes to origin independently. For a file requested from 50 regions, that’s 50 S3 GETs in a row. Origin Shield consolidates those into one fetch through a chosen regional cache.

We sized Origin Shield in us-east-1 (where most of the S3 origin traffic was anyway) and let CloudFront route edge misses through it. The team’s release shape — a new build downloaded by a million users globally over the first hour — is exactly the workload Origin Shield is built for.

Origin fetch rate dropped 73%. Effective cache hit rate (edge + shield) went from 84% to 96.4%. Net of Shield’s small operational cost, the S3 + egress bill dropped $22k/month. Edge latency improved 4% at p50 because the mid-tier was warmer.

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 →