Calculate network propagation delay from link distance and medium, plus RTT, transmission delay, bandwidth-delay product, and max distance for a latency budget.
Network Propagation Delay Formula
T_prop = D / (VF * c)
RTT = 2 * T_prop ; T_trans = L / R ; BDP = R * T_prop
Variables:
- T_prop is the one-way propagation delay in seconds
- D is the physical length of the link in meters
- VF is the velocity factor of the medium, the fraction of light speed at which the signal travels (0 to 1)
- c is the speed of light in a vacuum, 299,792,458 m/s
- RTT is the round-trip propagation delay, the theoretical floor for a ping time
- T_trans is the transmission delay, the time to push all bits of one packet onto the link
- L is the packet size in bits and R is the link rate in bits per second
- BDP is the bandwidth-delay product, the number of bits in flight on the link at once
The default mode applies the first formula directly: enter a distance, pick a medium, and the calculator multiplies the velocity factor by c to get the propagation speed, then divides distance by that speed. It reports the one-way delay, the round trip, and the delay per kilometer. The max distance mode rearranges the formula to D = T_prop * VF * c, so you enter a delay budget (one-way or round-trip) and get the farthest a link can reach within it. The propagation speed mode rearranges to VF * c = D / T_prop, which turns a distance and a measured delay into the implied speed and velocity factor and names the closest standard medium. The advanced fields in the default mode add the second set of formulas: packet size and link rate produce the transmission delay, the total one-way delay, and the bandwidth-delay product.
Propagation Speed by Medium and Delay Over Common Routes
The velocity factor is set by the physics of the medium, not by the quality of your internet plan. Radio waves in air move at essentially full light speed, while light in glass fiber and electrical signals in copper move at roughly two-thirds of it. The first table gives the values the calculator uses, plus the resulting delay per kilometer.
| Medium | Velocity factor | Speed (m/s) | One-way delay per km | Per 100 km |
|---|---|---|---|---|
| Wireless, radio, free space | 1.00 | 3.00 x 10^8 | 3.34 µs | 0.33 ms |
| Coaxial cable, foam dielectric | 0.83 | 2.49 x 10^8 | 4.02 µs | 0.40 ms |
| Optical fiber (single-mode) | 0.68 | 2.04 x 10^8 | 4.90 µs | 0.49 ms |
| Coaxial cable, solid polyethylene | 0.66 | 1.98 x 10^8 | 5.05 µs | 0.51 ms |
| Twisted pair, Cat5e/Cat6 | 0.65 | 1.95 x 10^8 | 5.13 µs | 0.51 ms |
| Textbook convention | 0.67 | 2.00 x 10^8 | 5.00 µs | 0.50 ms |
The second table applies the fiber rate of 4.90 microseconds per kilometer to real routes and compares the theoretical floor with round-trip times typically measured in practice. The gap is normal: submarine and terrestrial fiber paths run 20 to 50 percent longer than the great-circle line, and every router and amplifier along the way adds a little more. Use the gap, not the raw ping, to judge whether a path is efficient.
| Route | Great-circle distance | Theoretical one-way (fiber) | Theoretical RTT | Typical measured RTT |
|---|---|---|---|---|
| New York to Los Angeles | 3,944 km | 19.3 ms | 38.7 ms | 60 to 70 ms |
| New York to London | 5,585 km | 27.4 ms | 54.8 ms | 68 to 78 ms |
| San Francisco to Tokyo | 8,280 km | 40.6 ms | 81.2 ms | 100 to 120 ms |
| London to Singapore | 10,885 km | 53.4 ms | 106.8 ms | 160 to 180 ms |
| GEO satellite hop (radio, 35,786 km up and down) | 71,572 km path | 238.7 ms | 477.5 ms | 500 to 600 ms |
| LEO satellite hop (radio, about 550 km altitude) | 1,100 km path | 3.7 ms | 7.3 ms | 25 to 40 ms |
Example Problems
Example 1: One-way delay and RTT from New York to London over fiber.
The great-circle distance is 5,585 km and single-mode fiber has a velocity factor of 0.68. The propagation speed is 0.68 * 299,792,458 = 2.039 x 10^8 m/s. The one-way delay is T_prop = 5,585,000 / 2.039 x 10^8 = 0.0274 s, or 27.4 ms, and the round trip is 54.8 ms. A real transatlantic ping of about 70 ms is therefore only around 15 ms above the physical floor, which tells you the cable route is efficient.
Example 2: Total one-way delay for a packet, using the advanced fields.
A 1,500-byte packet crosses a 200 km fiber link at 100 Mbps. The transmission delay is T_trans = 12,000 bits / 100,000,000 bps = 120 microseconds. The propagation delay is T_prop = 200,000 / 2.039 x 10^8 = 981 microseconds. The total one-way delay is 120 + 981 = 1,101 microseconds, or about 1.1 ms. The bandwidth-delay product is 100,000,000 * 0.000981 = 98,100 bits, so about 12.3 KB of data is in flight on the link at any moment.
Frequently Asked Questions
Why is my measured ping higher than the calculated propagation delay?
The calculator gives the physical floor: the time the signal itself needs to cover the distance. A real ping adds four things on top. Cable routes follow roads, railways, and sea lanes, so the fiber path is usually 20 to 50 percent longer than the straight line. Every router adds processing delay while it inspects the packet. Congested links add queuing delay while the packet waits its turn. Finally, slow access links add serialization time at each end. If your ping is within roughly 30 percent of double the theoretical RTT, the path is about as direct as real infrastructure allows.
What propagation speed should I use for a homework problem?
Networking textbooks almost always state the speed in the problem, most commonly 2.0 x 10^8 m/s and sometimes 2.5 x 10^8 m/s. If the problem gives one, pick the matching textbook preset in the medium list so your answer matches the answer key. Only use the fiber or copper presets when the problem names a real medium without giving a speed, and use the wireless preset (full light speed) for satellite and radio links.
Does a faster internet plan reduce propagation delay?
No. Bandwidth sets the transmission delay, the time to push bits onto the link, but it does nothing to the time those bits need to travel the distance. Upgrading from 100 Mbps to 1 Gbps cuts the transmission delay of a 1,500-byte packet from 120 microseconds to 12, yet a New York to London round trip keeps its roughly 55 ms propagation floor. That is why latency-sensitive services put servers physically closer to users with CDNs and edge nodes instead of simply buying faster links.
