Networking · Performance
Network Latency Calculator
Work out how long a signal takes to cross a distance through fibre, copper, radio or a geostationary satellite, and the round-trip time (RTT) that sets the floor for any ping. Then see the bandwidth-delay product — the TCP window you need to fill a fast link — and how long a file really takes once handshakes and window limits are counted. It shows why a far-away server feels slow even on a fast connection.
Also listed inData & Storage
Your inputs
Quick start — try a typical case
Latency is how long one bit takes to arrive; bandwidth is how many bits arrive each second. Propagation delay finds the latency from distance; the other two show how latency and bandwidth together limit real downloads.
What the signal travels through. Light in glass fibre moves at about two-thirds of its speed in vacuum (≈ 204,000 km/s), so long-haul internet is slower than radio. A geostationary satellite adds 71,572 km of path — about 239 ms one way.
Straight-line (great-circle) distance between the two ends. Examples: Mumbai–Delhi ≈ 1,150 km, Mumbai–London ≈ 7,200 km, New York–London ≈ 5,570 km (3,460 mi).
Your result Live
One-Way Latency
4.9034 ms
1,000 km through optical fibre (glass)
- Round-Trip Time (RTT)
- 9.8068 ms
- Propagation Delay (one way)
- 4.9034 ms
- Signal Speed
- 203,940 km/s (68 % of c)
- Delay per 1,000 km
- 4.9034 ms
- Path Length Used
- 1,000 km
Working
One-way latency = path length ÷ signal speed + hops × per-hop delay
1: Step 1 · Signal speed in the medium
c ÷ 1.47 = 299,792.458 ÷ 1.47
Expression becomes 203,940.4 km/s
2: Step 2 · Path length
1,000 km × route factor 1
Expression becomes 1,000 km
3: Step 3 · Propagation delay
1,000 km ÷ 203,940.4 km/s × 1,000
Expression becomes 4.9034 ms
4: Step 4 · Processing delay at hops
0 hops × 0.1 ms
Expression becomes 0 ms
5: Step 5 · One-way latency
4.9034 ms + 0 ms
Expression becomes 4.9034 ms
6: Step 6 · Round-trip time (RTT)
2 × 4.9034 ms
Expression becomes 9.8068 ms
RTT = 9.8068 ms
The same 1,000 km path in each medium
4 rowsThis is the physical floor. A real ping is higher: cables rarely follow a straight line, and every router adds queuing, processing and serialisation delay.
To estimate a real-world ping, try a route factor of 1.3–1.5 (Advanced) and a few hops — or compare with what ping or traceroute shows.
ITU-T G.114 suggests keeping one-way delay under 150 ms for comfortable voice calls.
How this is calculated
One-way = distance × route factor ÷ (c × VF) + hops × per-hop delay · RTT = 2 × one-way · BDP = bandwidth × RTT · Max throughput = window ÷ RTT · Transfer time = (set-up RTTs + 1) × RTT + size ÷ min(bandwidth, window ÷ RTT)
- c
- Speed of light in vacuum, 299,792.458 km/s
- VF
- Velocity factor of the medium: fibre 1 ÷ 1.47 ≈ 0.680, copper pair 0.67, coax 0.66, radio 1.0
- RTT
- Round-trip time — there and back, in milliseconds
- BDP
- Bandwidth-delay product — data in flight, in bits or bytes
- window
- TCP receive window — data allowed in flight before an acknowledgement
Latency and bandwidth are different things. Bandwidth is how many bits a link carries each second (the width of the pipe); latency is how long one bit takes to get from end to end (the length of the pipe). Nothing travels faster than light, and in glass fibre light moves at about 204,000 km/s, so every 1,000 km adds about 4.9 ms each way. Mumbai to London is about 7,200 km in a straight line: at least 35 ms one way and 71 ms round trip before any routing, and real pings are higher. A far-away server feels slow even on a fast plan because each request, handshake and acknowledgement must wait a full round trip — a 2 MB web page that needs several round trips can take longer from London than from Mumbai on the very same 1 Gbps line. The bandwidth-delay product tells you how much data must be in flight to keep a link busy; if the TCP window is smaller, throughput is capped at window ÷ RTT no matter how fast the line is.
Assumptions & sources
- Fibre speed uses a refractive index of 1.47 (≈ 203,940 km/s). Copper twisted pair uses 0.67 c and coax 0.66 c; real cables vary by a few per cent — use a custom velocity factor from the datasheet if you have one.
- The geostationary hop assumes both ground stations sit directly under the satellite (35,786 km each way at the speed of light) — the shortest possible path.
- Transfer time ignores TCP slow start, packet loss and protocol overhead, and assumes the bandwidth is the bottleneck along the whole path.
- Set-up round trips: TCP handshake 1, TLS 1.3 adds 1, TLS 1.2 adds 2; one more round trip carries the request and the first byte back.
- 1Speed of light in vacuum, c = 299,792,458 m/s (exact) — NIST — CODATA Fundamental Physical Constants
- 2The International System of Units (SI Brochure) — definition of the metre — BIPM
- 3RFC 7323: TCP Extensions for High Performance (window scaling) — IETF / RFC Editor (2014-09). Obsoletes RFC 1323; explains the bandwidth-delay product and the 64 KiB window limit.
- 4RFC 1323: TCP Extensions for High Performance — IETF / RFC Editor (1992-05)
- 5RFC 8446: The Transport Layer Security (TLS) Protocol Version 1.3 — IETF / RFC Editor (2018-08). Full handshake in one round trip.
- 6ITU-T G.114: One-way transmission time — ITU-T. Recommends under 150 ms one-way delay for most interactive voice.
Calculator ID NET-LATENCY-001 · version 1.0.0
Common questions
What's the difference between latency and bandwidth?
Bandwidth is how much data a link carries per second; latency is how long each piece takes to arrive. Think of a motorway: bandwidth is the number of lanes, latency is the length of the road. Adding lanes doesn't make a 7,200 km journey shorter.
Why does a far-away website feel slow on my fast connection?
Loading a page takes many round trips — the TCP handshake, the TLS handshake, then each request. At 120 ms RTT, just opening an HTTPS connection and asking for the first byte costs 360 ms, whatever your plan speed. A server 5 ms away does the same in 15 ms. That's why sites use CDNs with servers near you.
How much latency does fibre add per kilometre?
About 4.9 µs per km one way (4.9 ms per 1,000 km), because light in glass moves at roughly 204,000 km/s. Round trip, double it: about 1 ms for every 100 km of cable.
Why is satellite internet so laggy?
A geostationary satellite sits 35,786 km above the equator. Up and down is 71,572 km, which takes about 239 ms even at the speed of light, so a request and its reply need about 477 ms — usually 600 ms or more in practice. Low-Earth-orbit satellites at around 550 km cut this to tens of milliseconds.
What is the bandwidth-delay product used for?
It sizes the TCP window and network buffers. A 100 Mbps link with 50 ms RTT holds 625,000 bytes (about 610 KiB) in flight. If the window is only 64 KiB, the connection tops out near 10.5 Mbps — about 10 % of the line — however fast the link.
Is my ping the same as the propagation delay?
No. Ping shows the round-trip time, which includes propagation both ways plus time in routers, queues and the servers themselves. Propagation delay is the physical minimum; real pings are often 1.5 to 2 times the straight-line fibre figure.