I know most of the network latency for short distances is due to router processing times. But for longer distances the speed of light also counts. And it's different from the speed the light in the vacuum. What is it?
-
related question http://serverfault.com/questions/61719/how-does-geography-affect-network-latency – Jader Dias May 29 '11 at 17:43
3 Answers
A typical index of refraction for optical fiber is 1.62, therefore the speed of light in a fiber is approximately 3e8 m/s / 1.62 = 1.85e8 m/s. Therefore it would take at least 1000000 m /1.85e8 m/s = 0.0054 s to travel that distance. Note that this value doesn't cover the extra distance traveled by the light from bouncing side to side.
- 111,361
- 10
- 201
- 247
-
1I expect more than this theoretical 5.4 ns/m, isn't there any source that says what actual latencies are? – Jader Dias May 28 '11 at 19:35
-
-
4@Jader, lol, first you ask for "minimum network latency" and then you complain they aren't realistic enough? You must be confused. – akappa May 28 '11 at 21:19
-
1@akappa I'm not confused. I'm asking the the "minimum network latency", not the "minimum latency through an ideal-but-yet-not-manufactured-medium" – Jader Dias May 28 '11 at 21:21
-
The fact that "fiber cables aren't laid in straight lines" has nothing to do with fiber technology. As for the actual speed of light in the medium, the informations you can find in the linked wikipedia article is more than adequate. – akappa May 28 '11 at 21:34
-
@akappa but not adequate enough to figure out the speed taking the internal reflections in account – Jader Dias May 29 '11 at 00:41
-
-
@akappa No they're not. Light speed is slower inside glass regardless of reflections. The existence reflections just mean that the light isn't following a straight path, even if the cable is straight. If I knew the average reflection angle for an optical fiber then I could calculate the average speed of a light beam inside it. – Jader Dias May 29 '11 at 01:04
-
From http://en.wikipedia.org/wiki/Refractive_index : The refractive index or index of refraction of a substance is a measure of the speed of light in that substance. It is expressed as a ratio of the speed of light in vacuum relative to that in the considered medium. – akappa May 29 '11 at 01:12
-
@Jader Dias: also take a look at that: http://en.wikipedia.org/wiki/Snell%27s_law , hope this clears your confusion – akappa May 29 '11 at 01:14
-
@akappa I thought in pointing the same articles to you. I hope you read them. Where the hell did you get the idea that the refractive index take in account the internal reflections? – Jader Dias May 29 '11 at 17:41
-
3@Jader Dias is right on that last point - the refractive index is a property of bulk material and can't take reflections into account. The actual delay factor due to these reflections depends on the angle of light rays entering the fiber. The maximum angle is determined by the NA of the fiber. Note that for thin (single-mode, less than 10 lambda diameter) fibers it doesn't even make sense to think about light rays bouncing around in the fiber... see http://en.wikipedia.org/wiki/Optical_fiber#Single-mode_fiber – Jonas Heidelberg May 29 '11 at 18:00
Distance Delay is simply the minimum amount of time that it takes the electrical signals that represent bits to travel down the physical wire. Optical cable sends bits at about ~5.5 µs/km, copper cable sends it at ~5.606 µs/km, and satellite sends bits at ~3.3 µs/km. (There are a few additional microseconds of delay from amplifying repeaters in optical cable, but compared to distance, the delay is negligible.)
source: http://www.networkperformancedaily.com/2008/06/latency_and_jitter_1.html
- 15,756
- 60
- 143
- 196
The general conception of the speed of light in vacuum is 300,000 km/s, give or take. That is 300 km/ms or around 1,000 km in 3.3 ms (milliseconds). The minimum network latency for a 1000 km connection using optic fibers may be between 10 and 30 milliseconds, according to the answer of @kyle kanos to the question "How fast does light travel through a fibre optic cable?", in the Physics Stack Exchange.
- 109
- 1
- 5