Network Speed and Transfer Time
Network plans are usually written in bit/s such as 100Mbps or 1Gbps, while file sizes are usually thought of in bytes. That mismatch is why line-rate numbers often feel disconnected from real transfer time. This page converts common line speeds into byte-based throughput, shows why measured throughput is lower than the theoretical maximum, and gives rough transfer-time estimates by data size. Unless noted otherwise, the examples below use decimal units (1MB = 1,000,000 bytes, 1GB = 1,000MB).
📶 Reading Network Speed Units
Start by connecting network notation to file-size notation.
bps vs B/s
- 🔤 Network links are usually labeled in bit/s such as Mbps or Gbps. File transfer feel is closer to byte/s such as MB/s.
- 🔤 1 byte = 8 bits, so to convert Mbps to MB/s you divide by 8.
- 🔤 For example, 100Mbps is theoretically 12.5MB/s and 1Gbps is about 125MB/s.
Common line speeds and their theoretical ceilings
| Link speed | Theoretical max | Rule of thumb |
|---|---|---|
| 10Mbps | About 1.25MB/s | Just over 1MB per second |
| 100Mbps | About 12.5MB/s | 100MB in about 8 seconds |
| 1Gbps | About 125MB/s | 1GB in about 8 seconds |
| 10Gbps | About 1.25GB/s | 10GB in about 8 seconds |
⚖️ Why Measured Throughput Is Lower
It is normal for real transfers to fall short of the advertised line rate. The useful question is where the gap comes from.
Main reasons throughput drops
- 🧩 TCP/IP and TLS add headers, acknowledgements, retransmissions, and other overhead, so not every bit is payload.
- 🧩 Wi-Fi is sensitive to distance, walls, interference, airtime sharing, and the number of connected devices.
- 🧩 Across the internet, the remote server, storage I/O, CPU, rate limits, and CDN path can all become the bottleneck.
- 🧩 Home internet is often asymmetric, so upload throughput may be much lower than download throughput.
Rough effective-throughput ranges
| Environment | Typical effective throughput vs theory | Notes |
|---|---|---|
| Wired LAN, good conditions | 70% to 95% | On a fast local network you can get quite close to the theoretical ceiling. |
| Wi-Fi, same room, light contention | 40% to 80% | Displayed link speed is often much higher than usable throughput. |
| Mobile network | Can vary from 10% to 70% or more | Congestion, movement, radio quality, and time of day matter a lot. |
⏱️ Transfer Time by Data Size
The table below shows how long transfers roughly take on 100Mbps and 1Gbps links. For the “measured” column, the estimates assume fairly good wired conditions: about 9 to 12MB/s on 100Mbps, and about 90 to 115MB/s on 1Gbps.
100Mbps vs 1Gbps
| Data size | 100Mbps theoretical | 100Mbps measured | 1Gbps theoretical | 1Gbps measured |
|---|---|---|---|---|
| 10MB | About 0.8s | About 0.9 to 1.1s | About 0.08s | About 0.09 to 0.11s |
| 100MB | About 8s | About 8.5 to 11s | About 0.8s | About 0.9 to 1.1s |
| 1GB | About 80s | About 1m 23s to 1m 51s | About 8s | About 8.7 to 11s |
| 10GB | About 13m 20s | About 14 to 19m | About 80s | About 1m 27s to 1m 51s |
How to use these estimates
- 📌 At around 100MB, a 100Mbps link feels like several seconds, while a 1Gbps link is around a second.
- 📌 Once transfers exceed 1GB, the difference between 100Mbps and 1Gbps becomes very noticeable.
- 📌 At 10GB scale, storage write speed and remote-side limits can matter as much as the network itself.
🧭 Practical Estimation Tips
Even rough estimates become useful if you keep the assumptions explicit.
Rules of thumb
- ✅ Treat link labels as decimal units. Start from 1Gbps ≈ 125MB/s.
- ✅ For rough planning, assume 70% to 90% of theory on wired links and 50% to 70% on Wi-Fi.
- ✅ Check upload speed separately when estimating uploads; download and upload are often not symmetrical.
- ✅ When transfers are slow, verify whether the bottleneck is the link, the remote server, or local storage.