CDN Cache Hit Savings Calculator

A cache hit serves a reusable response from a content delivery network (CDN) instead of asking the origin to send or rebuild it. At scale, that can remove a large share of origin data transfer and request processing. The financial value depends on what a hit actually avoids, not on the hit-rate percentage alone.

Two ratios matter. Cacheable share is the portion of traffic that may be stored and safely reused. Cache hit rate is the portion of that eligible traffic served from cache. A workload that is 80% cacheable with a 75% hit rate still sends all uncacheable traffic and one quarter of cacheable traffic to the origin.

Byte hit rate

Best suited to bandwidth-heavy workloads because it weights large responses by their transferred size.

Request hit rate

Useful when request fees, application work, database reads, or object-store operations dominate cost.

Miss amplification

Extra origin transfer caused by retries, revalidation, range requests, transformations, or multiple cache fills for one delivered response.

Cost comparisons also change with the decision being modeled. Improving an existing CDN normally leaves its delivery and request charges unchanged, so the comparison focuses on avoided origin work. Adding a CDN starts from direct-origin delivery and must include the new CDN charges as well as fixed operating and setup costs. A positive estimate is not permission to cache every response: personalized, authenticated, private, write-method, and deliberately bypassed traffic must stay outside the cacheable share unless the cache policy preserves the required separation. Provider rates should come from the same billing period, region mix, pricing tier, allowances, and discount structure as the traffic data.

How to Use This Tool:

Start with one representative monthly billing window and decide whether the comparison is an existing-CDN improvement or a new CDN deployment.

1. Choose Analysis mode. Use Optimize existing CDN when CDN delivery fees are already part of both cases. Use Add CDN in front of origin when the target introduces CDN delivery and request charges.
2. Enter Monthly delivered traffic, its unit, monthly requests, the safely cacheable share, and comparable current and target hit rates. TB and PB are decimal units; TiB equals 1,024 GB. A lower target models a regression rather than forcing a savings result.
3. Enter blended origin transfer and request costs. In add-CDN mode, also enter the target CDN rates. Use zero only when a cost is genuinely absent or immaterial.
4. Open Advanced when misses create extra transfer or the change has fixed monthly, setup, or planning-window costs. Review the monthly savings, break-even hit rate, and payback together before accepting the scenario.

Interpreting Results:

Net monthly savings is positive when the target total is lower than the baseline total and negative when the change costs more. Avoided origin transfer and requests explain where the gross reduction comes from; they do not include CDN or fixed add-on charges.

• A break-even hit rate from 0% through 99% is the minimum modeled target that covers the added monthly cost.
• A break-even result at or below 0% means the modeled target costs are covered without a positive cacheable hit rate. A result above 99% means hit-rate improvement alone cannot recover them within the supported range.
• Payback is available only when setup cost is greater than zero and monthly savings is positive.
• Verify the chosen hit-rate metric against invoices and origin telemetry. Mixing byte hit rate with request-based cost assumptions can make a precise-looking estimate misleading.

Technical Details:

The model divides delivered traffic into cacheable and uncacheable portions, then sends every uncacheable byte and request to origin. Only the miss fraction of cacheable traffic reaches origin. The same cacheable share and hit-rate assumption are applied to bytes and requests, so use inputs that make that simplification reasonable.

Formula Core:

For total monthly traffic T in GB, monthly requests R in millions, cacheable fraction c, hit fraction h, and miss factor m, target origin load is:

Miss amplification is entered as a percentage, so m = 1 + amplification ÷ 100. It affects origin transfer but not the request count. In add-CDN mode, the direct-origin baseline is Tm GB and R million requests.

Let e be origin USD/GB, q be origin USD per million requests, and D be all target-only CDN and fixed monthly charges. Monthly savings is the baseline cost minus the target cost:

Annual savings equals 12S, planning-period savings equals S multiplied by the selected number of months, and payback months equals setup cost divided by S when S is positive.

Break-even rule:

The value of a 100-percentage-point improvement across cacheable traffic is V = Tcme + Rcq. With unavoidable origin cost U, the break-even hit rate is:

The result is reported as a percentage. It is unavailable when cacheable traffic has no modeled origin value. Inputs are calculated at full precision; currency, percentage, and payback displays may round for readability.

Limitations and Safety Notes:

• The estimate assumes one cacheable share and one hit rate can represent both traffic volume and requests. Model separate workload classes when their sizes, request costs, or cache behavior differ materially.
• Tiered rates, included allowances, regional transfer, origin shield charges, invalidations, logging, support plans, and compute savings belong in the blended rates or fixed cost when applicable.
• Higher hit rates are not automatically safe. Cache keys, authorization rules, cookies, tenant boundaries, response directives, and purge behavior must prevent private or stale content from being reused incorrectly.
• Presets are editable examples, not current provider quotes or cache-policy recommendations.

Worked Examples:

References:

• Cache performance, Cloudflare, May 7, 2026.
• How to enable caching in Nginx, Simplified Guide.