Cloud egress: Cloudflare R2 vs Backblaze B2 compare

Path matrix: Cloudflare R2 vs Backblaze B2 download and ops topics (educational)
Topic	R2	B2	Check
Internet download pricing	Often $0 egress (verify)	Free egress allotment vs storage; overage rules	Current ToS + pricing
CDN pairing	Cloudflare ecosystem affinity	Partner CDN programs	Contract specifics
Minimum storage / retention	Product rules evolve	Storage minimums may apply	Bill surprises
S3-compatible API ops	Class A/B ops style pricing	API call pricing	Chatty workloads
Egress to other clouds	Confirm exceptions	Download to AWS/GCP may differ	Destination path

Path matrix: Cloudflare R2 vs Backblaze B2 download and ops topics (educational)

Topic

Check

Internet download pricing

Often $0 egress (verify)

Free egress allotment vs storage; overage rules

Current ToS + pricing

CDN pairing

Cloudflare ecosystem affinity

Partner CDN programs

Contract specifics

Minimum storage / retention

Product rules evolve

Storage minimums may apply

Bill surprises

S3-compatible API ops

Class A/B ops style pricing

API call pricing

Chatty workloads

Egress to other clouds

Confirm exceptions

Download to AWS/GCP may differ

Destination path

Provider source footnotes

Cloudflare R2 pricing — as-of 2026-09-09 Unknown until verified

Backblaze B2 pricing — as-of 2026-09-09 Unknown until verified

Backblaze B2 bandwidth / partner notes — as-of 2026-09-09 Unknown until verified

Dated starting points only — Unknown until verified means no invented $/GB. PipeToll does not scrape private consoles or guarantee figures.

This spoke compares object-store egress path classes with path templates. Rates stay Unknown until verified. Do not treat “free egress” as universal. Caps and excluded destinations appear in fine print. PipeToll compares path templates and allotment shapes — not logo preference.

Path templates to model

Browser ← Internet ← bucket — public asset download; watch allotments vs stored bytes.
Browser ← CDN ← bucket — who bills origin fill vs edge bandwidth?
AWS/GCP job ← pull from B2/R2 — multi-cloud download exceptions may apply.
App server chatty LIST/GET — ops pricing can dominate small-object workloads.
Backup restore spike — one-month download multiples stress free-egress caps.

Common traps

Annualizing a quiet month understates restore spikes. Ignoring Class A/B or API call fees undercounts listing-heavy apps. Assuming partner CDN programs match public list prices is unsafe. No savings guarantees when switching vendors — contracts and migration egress both matter.

FinOps checklist

Measure stored GB and downloaded GB for 90 days; note peak restore months.
Read current free-egress multiples and excluded destinations.
Separate ops/request cost from download GB in the model.
Run sensitivity in the calculator (?mode=gb&path=r2-b2); keep Unknown badges honest.
Cross-check S3 vs R2 and glossary terms.

FinOps-lite Unknown until verified

Worked example (educational): 3× storage download month

If you store 10 TB and download 30 TB in a month, allotment rules (e.g. free egress multiples of stored bytes) may flip from $0-looking to overage. Model download GB × overage $/GB in Mode A only after reading current caps. Treat public blog numbers as Unknown until verified. No guaranteed savings versus R2 or S3.

FAQ on this path

Do free egress claims cover multi-cloud pulls?

Not always. Destination exceptions and partner CDN rules appear in fine print. Verify before migration math.

What breaks a “cheap egress” story?

Chatty LIST/PUT ops, retention minimums, and download caps relative to stored bytes.

How should FinOps compare R2 and B2?

Normalize path templates and ops classes first; then apply dated rates — never crown a winner from Incomplete cells.