Byte
Eight bits in this calculation.
Unit Converters
Keep bytes, bits, decimal prefixes, binary prefixes, nominal link rate, and payload efficiency explicit in one transfer planning record.
Digital units and timing
Convert one payload across byte units, distinguish decimal from binary prefixes, and calculate ideal and efficiency-adjusted transfer duration from a bit-rate link.
| Representation | Value | Unit | Prefix basis |
|---|
How to use
Convert one payload across byte units, distinguish decimal from binary prefixes, and calculate ideal and efficiency-adjusted transfer duration from a bit-rate link.
Eight bits in this calculation.
kB, MB, GB use powers of 1000.
KiB, MiB, GiB use powers of 1024.
Nominal bits transmitted per second.
Fraction of line capacity carrying the declared payload.
Result interpretation
The ideal time assumes every transmitted bit is payload. The adjusted time divides nominal rate by the entered efficiency, exposing protocol and operational loss without pretending to model every network effect.
Calculation method
Convert selected size to bytes, multiply by eight to obtain bits, normalize rate to bit/s, divide bits by nominal rate for ideal time, then multiply rate by efficiency and divide again for adjusted time.
Evidence controls
Operating systems, cloud products, storage vendors, and applications may label the same byte count differently.
Decide whether the size includes archives, headers, encryption, checksums, retries, replication, or only application content.
Do not confuse bits per second with bytes per second.
Use measured goodput or a documented planning ratio rather than a convenient universal percentage.
End-to-end time is limited by the slowest relevant path, storage, CPU, or application stage.
One-time setup, handshake, queue, and verification delays are outside this reference formula.
Concurrent flows may improve utilization or divide capacity; this single-flow model does not predict either.
Visual explanation
The unit ledger starts from one exact byte count. The paired timelines show the minimum line-rate duration and the longer payload-efficiency duration on the same scale.
Detailed calculation process
bits = 8B; tideal = bits/R; Reffective = Rη; teffective = bits/Reffective
| Symbol | Meaning | Required unit |
|---|---|---|
| B | payload size | bytes |
| bits | payload bit count | bit |
| R | nominal line rate | bit/s |
| η | payload efficiency | 0 to 1 |
| tideal | ideal line-rate time | s |
| teffective | efficiency-adjusted time | s |
Reconciliation:Waiting for current inputs.
Defaults and assumptions
Defaults demonstrate 10 decimal GB over a 100 Mbit/s link at 85% payload efficiency. They are not a speed-test result or service guarantee.
| Check | Current value A | Current value B | Decision role |
|---|
Decision analysis
Use ideal time as a physical lower-bound comparison and adjusted time as a transparent planning estimate. Replace efficiency with evidence from the actual protocol, endpoints, path, and workload.
First reconcile the byte count: a 10 GB vendor payload and a 10 GiB file are not equal. Then confirm rate units; a 100 MB/s storage figure is eight times a 100 Mbit/s network figure before overhead. If planning operational windows, add setup, queueing, retry, verification, encryption, compression, and storage time separately. Effective payload rate should be measured over enough data that startup latency does not dominate, while small-object workloads require a different model.
Verification workflow
Obtain the payload byte count from a manifest, object metadata, or file system rather than relying only on a rounded display label. Decide whether directories, metadata, sparse extents, archive headers, checksums, encryption expansion, and replicas are inside the scope. Record both exact bytes and the displayed decimal or binary unit. When comparing vendor capacity with operating-system display, reconcile the same physical bytes before calling the difference loss or overhead.
A physical line rate, interface throughput, IP throughput, transport throughput, application goodput, and storage write rate describe different boundaries. The efficiency ratio must connect the entered nominal rate to the declared payload bytes at a compatible boundary. Do not apply protocol overhead to a goodput measurement that already excludes it. Record direction, duplex behavior, and whether the service rate is shared, burstable, shaped, or guaranteed.
Use sufficiently large representative transfers to reduce startup dominance, or model startup separately for small objects. Capture endpoints, path, protocol, encryption, compression, concurrency, packet loss, retransmissions, CPU, and storage state. Calculate useful payload bytes divided by elapsed time, then divide that goodput by comparable nominal bit rate. Report a range or percentile when conditions vary rather than presenting one optimistic speed-test observation as universal efficiency.
Add connection setup, authentication, queueing, scanning, compression, encryption, read, write, checksum, verification, unpacking, and retry time when planning a maintenance or recovery window. Identify which stages overlap and which are sequential. The ideal card is a lower-bound comparison; the efficiency-adjusted card remains a bulk-transfer estimate. Small-object workloads and high-latency paths commonly need per-object or protocol-window models instead of one proportional calculation.
After transfer, compare exact delivered bytes, elapsed timing boundary, and measured goodput with the estimate. A slower result can reflect network contention, storage, CPU, shaping, retransmission, or payload-scope mismatch. A faster result may indicate compression, cache, deduplication, or an incorrect source byte count. Preserve the planned assumptions and observed evidence separately. Update future efficiency inputs only from comparable controlled observations, not by editing the old calculation to fit reality.
Evidence and data lineage
Retain exact bytes, displayed label, prefix convention, nominal service rate, direction, interface counters, measured goodput method, efficiency derivation, protocol, endpoints, path, concurrency, timestamp, and workload definition.
Limits and exclusions
The page does not simulate TCP behavior, congestion, packet loss, latency, startup, retries, compression, encryption, storage I/O, queueing, checksums, parallel flows, or service-provider shaping.
Reference framework
The International System of Units defines decimal prefixes such as kilo, mega, and giga, while IEC binary-prefix conventions distinguish kibi, mebi, and gibi for powers of 1024. IETF RFC 6349 provides a framework for measuring TCP throughput and emphasizes test-path and host considerations beyond nominal bandwidth. This calculator uses exact byte factors, eight bits per byte, and an entered efficiency ratio; it does not reproduce a protocol performance model. Cite exact object metadata, interface or service definitions, and the measurement procedure used to derive goodput. Prefix labels and layer boundaries must remain explicit whenever timing supports capacity or recovery planning. When a storage or network contract uses an ambiguous abbreviation, quote its formal definition rather than assuming capitalization alone resolves it. For audit work, preserve the raw integer byte count and integer interface counters so later unit formatting cannot alter the underlying evidence. Record rounding only at the presentation boundary, with the exact value retained for every subsequent reconciliation and decision.
Worked decision cases
Exact archive bytes and measured goodput produce a better window estimate than capacity labels alone.
Decimal storage size and binary operating-system display are reconciled before claiming missing capacity.
Important note
Always name both the size prefix and the rate unit. GB, GiB, MB/s, and Mbit/s are not interchangeable labels.
The link rate is entered in bits per second.
No. MB is 1,000,000 bytes; MiB is 1,048,576 bytes.
No. Uppercase B means bytes; lowercase b means bits.
Every nominal line-rate bit would carry declared payload, an idealization.
Yes, when its evidence is defined consistently with payload and line rate.
Effective payload rate is nominal rate multiplied by efficiency.
Yes, especially for small or repeated transfers, but it is outside this reference page.
Yes if bytes sent shrink, but compression CPU and ratio need a separate model.
Quote exact bytes plus the chosen display unit and prefix convention.
Only as nominal input; real payload efficiency can vary greatly.