Calculator
AdjustedBps = (SpeedBps × BondedLinks) × Utilization × (1 − Overhead)
TCPcapBps = (TCPwindowBytes × 8) ÷ RTTseconds
EffectiveBps = min(AdjustedBps, TCPcapBps)
TimeSeconds = (EffectiveDataBytes × 8) ÷ EffectiveBps + SetupSeconds
- Enter your data size and pick the correct unit.
- Set link speed, utilization, and protocol overhead.
- Add compression and multiplier if your workflow changes size.
- Optionally apply RTT and TCP window for long-distance links.
- Press Calculate to see transfer and total durations.
Example data table
| Scenario | Data | Speed | Util / Overhead | Compression | Estimated time |
|---|---|---|---|---|---|
| Home upload | 25 GB | 20 Mbps | 70% / 6% | 1.00× | ≈ 4h 14m |
| Office backup | 500 GB | 1 Gbps | 85% / 5% | 1.25× | ≈ 1h 15m |
| Cloud migration | 4 TB | 500 Mbps | 80% / 8% | 1.50× | ≈ 19h 09m |
| Datacenter link | 12 TB | 10 Gbps | 90% / 3% | 1.00× | ≈ 3h 01m |
Estimates vary with congestion, retransmits, storage speed, and server limits.
Notes and engineering tips
- Measure real throughput with a test transfer.
- Set utilization to your 95th percentile rate.
- Increase overhead if you see many retries.
- Add setup time for batch jobs and staging.
Units and data size conversion
Accurate transfer forecasts start with consistent units and realistic inputs. The calculator converts file size using binary (1024) or decimal (1000) steps, then interprets network rates in bps, Kbps, Mbps, and Gbps. A 10 GB dataset is 10,000,000,000 bytes in decimal, but 10,737,418,240 bytes in binary. That gap can shift hours on large jobs, so select the unit system that matches your reporting source.
Effective bandwidth with utilization and overhead
Throughput is rarely equal to the advertised link rate. Utilization represents the sustainable share of the link, while protocol overhead models headers, framing, encryption, and retransmits. For example, 100 Mbps at 80% utilization and 5% overhead yields 100×0.80×0.95 = 76 Mbps effective. On shared networks you may set utilization to 40–60% to reflect contention, shaping, or throttling.
Latency, RTT, and TCP window limits
Latency affects long‑distance transfers even when bandwidth is high. If you set RTT and a TCP window size, the calculator applies a throughput cap: windowBytes×8 / RTTseconds. With a 1 MB window and 200 ms RTT, the cap is about 41.9 Mbps, regardless of a faster line. Larger windows, parallel streams, and tuned congestion control can raise throughput on high‑delay paths.
Compression and workflow multipliers
Data moved over the wire may differ from the original size. Compression factor reduces transferred bytes when the payload is compressible, while the multiplier increases bytes for replication, parity, checksums, or container overhead. A 2.0 compression factor halves transferred data, but only when content compresses well. A 1.20 multiplier adds 20% for copies or redundancy. Use measured ratios from prior runs to avoid optimistic planning.
Operational planning and report exports
The calculator reports data‑only time and total time by adding setup seconds for staging, authentication, or job startup. Bonded links multiply bandwidth only when aggregation is real end‑to‑end; otherwise, keep links at one and lower utilization. Export CSV for logs or PDF for approvals. Add a contingency for verification, retries, and integrity checks. Pair the estimate with storage write speed and API limits, because endpoints bottleneck before the link does. For cutovers, schedule total time inside a maintenance window, then communicate assumptions to stakeholders early and clearly.
FAQs
Why does my estimate differ from real transfers?
Real throughput depends on congestion, packet loss, server limits, and disk read/write speed. Calibrate utilization and overhead using a test run, then add setup time for staging and verification.
Should I choose binary or decimal units?
Use binary when sizes come from operating systems and storage tools. Use decimal when sizes come from marketing specs or network plans. Staying consistent prevents silent errors when estimating multi‑terabyte transfers.
What overhead value should I use?
Start with 3–8% for clean Ethernet links. Use 8–15% when tunneling through VPNs or when encryption adds headers. Increase overhead when you observe frequent retransmits, small packets, or unstable wireless.
How do RTT and TCP window affect speed?
If both are set, throughput is capped by windowBytes×8 divided by RTTseconds. High latency with a small window can limit fast links. Larger windows, window scaling, and parallel streams reduce this constraint.
How do I model compression accurately?
Run a representative sample transfer and compare sent bytes to original bytes. Set compression factor to originalBytes divided by sentBytes. Use 1.00 for already‑compressed media, encrypted archives, or random data.
Can bonded links always multiply bandwidth?
Only when aggregation is supported end‑to‑end and traffic can be distributed across links. If sessions stick to one path, gains may be small. In that case, keep links at one and reduce utilization to match reality.