Worth investigating when…
- The same exact documents recur frequently.
- Restored content must match the original.
- At least part of storage spending can fall with usage.
01 / Cloud Cost & FinOps
Same documents.
Less repeated storage?
If your service stores and retrieves the same reference material again and again, repetition may be a useful place to investigate.
Explore one controlled test of public HTTP standards documents, then assess whether it fits your work.
See whether your workload fits Recorded evidenceRFC / REPEAT-HEAVY
Same test workload.
Different storage footprint.
Normalized to the comparison method. Not actual GB or a prediction for your data.
98.1%less storage
9.5×faster end-to-end replay
Reported for this specific repeat-heavy test. Not a reduction in your total cloud bill.
Source: public_corpus_suite_report.md, “Public RFC retrieval hotset.” Archived supplied results, not a live test or an independently reproduced result.
02 / Put the result in context
A promising pattern.
Not a blanket promise.
Repeat-heavy data is the point of this example. The next decision is whether your workload has the same useful characteristics.
Check the limits first.
- Data with little repetition may not benefit.
- A fixed bill can stay fixed even when usage falls.
- Requests, transfer, integration and operating costs still matter.
The source includes a counterexample.
Its low-repeat public control used 26.8% more storage and ran at 0.8× baseline replay speed. That is a separate scenario, not the RFC result above. Repetition and workload selection matter.
03 / Optional cost exampleCould lower storage usage change the bill?Try an illustrative estimate after reviewing the evidence.
Illustrative monthly storage bill
Before$149.85
With the assumed reduction$2.70
Difference before additional costs$147.15 / month
After additional costsNot calculated
Enter additional monthly costs to estimate net savings.
This is one storage component, not total cloud spending. Annual figures assume 12 equal months. Negative savings mean a cost increase.
How the estimate is calculated
Retained reference: Cloudflare R2 Standard storage at $0.015 per GB-month, with 10 GB-month free and whole-unit rounding. Checked in the saved suite on September 21, 2026. View provider pricing.
- Apply the reported 98.1% reduction to the entered volume.
- Round each volume up to a whole GB-month, subtract the 10 GB-month allowance, and multiply billable units by $0.015.
- Apply the entered minimum payment. If the bill cannot decrease, keep the “after” bill equal to the “before” bill.
- Subtract known additional monthly costs. Leave net savings unknown if those costs are blank.
Request and retrieval operations, data transfer, taxes and other costs are not automatically included. All copies, indexes and metadata must fit within the measured storage boundary. This scenario has not been tested on R2.
Evidence & limitationsWhat sits behind these numbers?
Recorded scenario
- Workload
- Public RFC retrieval hotset: real HTTP standards texts repeatedly replayed in a documents/retrieval workflow.
- Comparison
- A basic exact-storage and replay path, not a best-of-market storage comparison.
- Source
public_corpus_suite_report.mdin the supplied AxiomARC v6.2 proof bundle.- Date and hardware
- August 7, 2026 is the archive date. The scenario’s exact run time and hardware configuration are not stated in the archived summary.
Source-reported results
| Storage reduction | 98.1% |
|---|---|
| End-to-end replay speed | 9.5× baseline |
| Retrieval p95 speed | 21.1× baseline |
| Deduplication ratio | 63.0× |
| Verification | Reported passed |
“p95” describes the slow-end threshold covering 95% of requests. Source-reported verification is not independent reproduction of the test.
No live AxiomARC engine runs on this page. The test does not establish customer performance, production scaling, total cloud savings, or the commercial terms of an OMporium service.