Payload-throughput duration surfaceLogarithmic color progression; exact time printed in every cell
Unit Converters
Data Transfer Range Calculator
Combine payload and throughput uncertainty into a four-corner transfer-time envelope, preserving the optimistic, pessimistic, and mixed cases so planning ranges remain traceable.
TRANSFER-TIME ENVELOPE
Pair payload and throughput endpoints without losing the corner cases
For migration and maintenance-window planning, this calculator converts a declared payload interval and sustained-throughput interval into four traceable durations. It does not assign probabilities to the endpoints.
Corner rule: shortest time uses minimum payload with maximum throughput; longest time uses maximum payload with minimum throughput. Reversed endpoints are rejected rather than silently sorted.
Shortest duration—
Longest duration—
Envelope width—
Worst / best ratio—
Payload interval—
Throughput interval—
DURATION SURFACE
Read time across both uncertain inputs at once
Rows move from smaller to larger payloads; columns move from faster to slower throughput. Every cell is a duration, so the diagonal exposes how jointly adverse assumptions widen the window.
| Payload endpoint | Throughput endpoint | Duration | Role |
|---|
HOW TO USE
Define a defensible transfer window
- Inventory the smallest and largest payload scope that could enter the migration.
- Use sustained low and high throughput from representative end-to-end tests, not interface nameplate speed.
- Keep decimal GB and binary GiB distinct at both payload endpoints.
- Use 100% efficiency when throughput already measures payload goodput; otherwise document the derating factor.
- Read all four corners and state which scenario the approved maintenance window protects.
RANGE FUNDAMENTALS
Six quantities define the envelope
CALCULATION METHOD
Time increases with bytes and decreases with rate
t(P,v) = 8P / (vη)tbest = 8Pmin / (vmaxη)tworst = 8Pmax / (vminη)Δt = tworst − tbestP is bytes, v bits per second, η dimensionless efficiency, and t seconds.
ENDPOINT DISCIPLINE
Do not reorder evidence to make the range convenient
The calculator rejects a maximum below a minimum. Automatic sorting can conceal mislabeled evidence or a unit error. Freeze whether snapshots, replicas, sparse regions, deleted-object retention, and encryption overhead belong inside each endpoint.
THROUGHPUT CONDITIONS
Low and high rates must describe the same path boundary
A low rate from an encrypted WAN test and a high rate from a local unencrypted copy do not form a coherent interval. Match protocol, concurrency, storage systems, encryption, compression, and measurement window.
WINDOW POLICY
The worst corner is conservative only for modeled variables
Maximum payload with minimum throughput protects against both entered endpoints. It says nothing about outages, retries, API quotas, manifest scans, validation, or cutover. Add those operational times separately and name their owner.
DETAILED CALCULATION PROCESS
Default four-corner substitution
| Symbol | Meaning | Default | Canonical value |
|---|---|---|---|
| Pmin | Minimum payload | 100 GB | 100,000,000,000 B |
| Pmax | Maximum payload | 500 GB | 500,000,000,000 B |
| vmin | Minimum rate | 1 Gbit/s | 1,000,000,000 bit/s |
| vmax | Maximum rate | 10 Gbit/s | 10,000,000,000 bit/s |
| η | Payload efficiency | 0.80 | dimensionless |
- Convert GB endpoints to bytes with powers of 1,000.
- Convert Gbit/s endpoints to bits per second.
- Best: 8 × 100 GB ÷ (10 Gbit/s × 0.80) = 100 s.
- Small/slow: 8 × 100 GB ÷ (1 Gbit/s × 0.80) = 1,000 s.
- Large/fast: 8 × 500 GB ÷ (10 Gbit/s × 0.80) = 500 s.
- Worst: 8 × 500 GB ÷ (1 Gbit/s × 0.80) = 5,000 s.
- Envelope width = 5,000 − 100 = 4,900 s.
- Reverse check: worst/best = (500/100) × (10/1) = 50.
RESULT INTERPRETATION
A wide ratio identifies sensitivity, not probability
A large ratio means data scope and path performance jointly matter. The surface locates which assumption drives duration but cannot tell which cell is likely. Probability requires distributions and dependence evidence.
EVIDENCE AND LINEAGE
Keep every endpoint reproducible
Retain inventory query, timestamp, object boundary, unit, source and target systems, protocol, concurrency, encryption state, path, test duration, achieved payload bytes, nominal capacity, and the authority for efficiency. State whether validation and cutover sit outside the transfer time.
LIMITS AND EXCLUSIONS
What the envelope does not contain
- No probability attaches to cells or corners.
- Throughput is constant within each scenario.
- Startup, retry, verification, and cutover are excluded.
- Payload growth during transfer is not modeled.
- Correlated endpoints require separate scenarios.
WORKED DECISION CASES
Different policies use different corners
Overnight migration
The team protects the maximum-payload/minimum-rate corner because downtime is costly, then adds validation and cutover as separately scheduled work.
Background replication
The middle cells are acceptable, but the worst corner violates recovery lag. The team constrains admitted payload rather than assuming the high rate.
TECHNICAL GLOSSARY
Terms used by the duration surface
- Endpoint
- A documented lower or upper bound.
- Corner case
- One endpoint combination.
- Goodput
- Useful payload per second.
- Envelope
- Minimum-to-maximum modeled result.
- Deterministic range
- A bound without probability.
- Rate limiter
- A resource or policy capping flow.
- Maintenance window
- Reserved operational time.
- Correlation
- Dependence that changes valid pairings.
IMPORTANT NOTE
Do not label this interval “95%” or “expected”
The calculator evaluates declared limits only. A probabilistic claim requires distributions, dependence assumptions, and evidence beyond endpoint arithmetic.
Frequently asked questions
Why is best paired with maximum rate?
Time is inversely proportional to throughput.
Should efficiency be 100% for measured goodput?
Yes, when the rate already measures useful payload.
Can minimum payload be zero?
Yes, but the worst/best ratio becomes undefined.
Why reject reversed endpoints?
Sorting could conceal mislabeled evidence.
Does the surface include latency?
Only insofar as sustained-rate tests already contain its effect.
Can I add a safety buffer?
Add it as a separately documented operational allowance.
Are binary units supported?
Yes; select the actual payload unit.
What if throughput changes over time?
Use segmented intervals or a time-series model.
AUTHORITATIVE BASIS