6

so latency is the time it takes for bits to get from the sender to the receiver and the speed is how fast bits travel , so logically the more speed the network has the more low latency it has also , but why are people saying that Satellite connection has good speed but Horrible Latency !! can someone please explain what going on here ?

10

You can have high speed, but if you have to go through a long distance then you have also high latency.

Let's use Los Angeles and London to understand this:

The distance from Los Angeles to London is around 9,000 kilometers.(The shortest path).

A geostationary satellite is 35,800 kilometers above the earth.

The satellite is roughly 4 times farther than the LA-London distance.

As radio waves go at the speed of light, which is 300,000 km per second, when you are communicating with a satellite the total distance (going up and down) is nearly 71,500 kilometers. The time that the radio wave takes to move from origin to destination will be around 240 ms (milliseconds).

Using fiber optic the speed is around 2/3 of the speed of light (200.000 km per second), and as the distance is around 9,000 kilometers the travel time is near 45 ms.

Satellite is useful for one-way communication as TV, but not suitable for two-way communication as video-conference.

| improve this answer | |
  • jcbermu Thank you so much for your time and explaining this to me ! now i understand. – justin Dec 15 '17 at 12:05
7

Latency is really just distance: the time it takes for the beginning of a bit to leave here and arrive there. (As perfectly explained in the other answer) this is entirely unrelated to speed: it takes approx 240 ms for the signal to start arriving on a satellite link, whether it's going at 1 bit/second or 1 Mbit/second. The beginning of the message takes 1 lag period to arrive. The end of the first bit takes 1 lag + 1 bit time. If you're waiting for an answer, it usually takes 1 lag + N bit times out, 1 processing time P and 1 lag + M bit times for the answer (N and M being message and answer bit lengths) = 2 * lag + (N + M) bit times + P. Note that lag is often much bigger than packet time, and often dominates the transaction.

I usually use the example of a small horse. It goes approx 20 metres/sec and it's approx 2 metres long. Racing over a 1 km, it takes 1000/20 = 50 sec for the nose to arrive, and the tail arrives 2/20 = 0.1 sec afterwards. If you change the distance of the race, the tail still arrives 0.1 sec after the nose.

The analogy breaks down very slightly because the speed of propagation and bit time are basically unrelated in electrical and optical signals. The propagation is extremely fast and pretty uniform (ranging from approx 0.6 c to nearly 1 c) and so latency only depends on distance. Whereas the bit rate varies enormously (as it costs money to engineer speed and reliability). Non-specialised speeds in common commercial use are as low as 9600 bit/sec and as high as 10 Gbit/sec.

Note that measuring propagation delay can be tricky, and often we measure from the beginning of transmission to the end of reception, ie the delay from issuing the order to transmit to the beginning of the intended action.

A proper understanding of the importance of lag tells you why

  • We have Sliding windows in TCP
  • We need Buffering for audio and video
  • We need Client-server architecture for high level protocols eg databases
  • We can't use TFTP over long-lag lines
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.