Skip to main contentSkip to user menuSkip to navigation
Interactive design lab

CDN Capacity & Failure Planner

How will a CDN behave when traffic, cache reuse, or a delivery path changes?

A content delivery network serves reusable responses from edge locations near users. Its value comes from shortening the hit path and preventing cache misses from overwhelming the origin.

Step 1

Model

Edit traffic, payload size, cache hit rate, edge and origin capacity, regional footprint, shield behavior, stale coverage, and unit costs. Every number is an illustrative assumption that you can replace.

Step 2

Observe

Trace requests through the edge, optional shield, and origin. Compare origin offload, p95 and p99 latency, availability, egress, saturation, and monthly run-rate.

Step 3

Challenge

Start with normal delivery, then inject a global purge, low-hit workload, regional loss, or origin outage. Adjust the plan until both the request path and capacity envelope recover.

Editable CDN planning model

Trace every request from edge to origin

Change demand, cache behavior, routing, capacity, and unit costs. The figures are illustrative assumptions, not vendor quotes or performance claims.

Challenge the healthy path

Plan is operating under pressure

Tail latency or capacity headroom is outside the recommended planning range.

Scenario: Normal delivery

Origin offload

93.0%

1,837 of 26.4K req/s reach origin

p95 latency

259ms

Fast path unless more than 5% of requests miss

p99 latency

313ms

Tail follows the miss or failure path

Availability

100.00%

0 req/s cannot be served

Monthly egress

12.3PB

857.3TB leaves the origin

Monthly run-rate

$533,033

Selected conditions projected for 30 days

Live request map

Where the peak load is served

26.4K incoming req/s · 88% effective hit rate

Users

26.4K req/s

2.2x peak demand

3 edge regions

23.2K hits

80% capacity used

Origin shield

1,331 reused

42% of edge misses

Origin

1,837 req/s

57% capacity used

Request disposition

Served at edge23.2K req/s

88.0% of incoming demand

Reused at shield1,331 req/s

Misses absorbed before origin

Served by origin1,837 req/s

1,837 req/s consume origin capacity

Failed or rejected0 req/s

0 req/s are not served

Capacity envelope

The marker indicates 100% physical capacity. Plan below 75% to preserve recovery headroom.

Edge saturation80%

3 active × 11K req/s per region

Origin saturation57%

1,837 demand / 3,200 req/s capacity

Editable cost model

What creates the monthly run-rate

These inputs are illustrative planning assumptions. Replace them with your own contract, region, and workload values.

Edge delivery

$431,101

12.3PB × $0.04/GB

Origin egress

$64,296

857.3TB × $0.08/GB

Request processing

$37,636

68.4KM × $0.55

Planning consequence

The modeled path serves 100.00% of peak demand with 93.0% origin offload. Preserve the remaining headroom for deploys, retries, and regional failover.

Effective hit 88%Edge headroom 20%Origin headroom 43%
No quiz questions available
Could not load questions file
Spotted an issue or have a better explanation? This page is open source.Edit on GitHub·Suggest an improvement