Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Suppose i had a fiber optic link from one side of the planet to the other side of the planet.

Is it safe to say, that with current technology, the latency of communication can never be reduced?

Understand that a fiber optic cable is not a perfect medium thus data only travels close to the speed of light.

Also lets consider that i will not be drilling a hole thru the center of the earth and it is just running along the ocean and or land.

Update: Thank you all who responded. It's a little saddening that current technology has this limitation. In the mean time I will consider CDNs and some sort of geo load balancing. Cheers!

share|improve this question
Jerry Mayers raised an interesting point that the protocol matters too. But baring that point for now and just considering current implementations. serverfault.com/questions/134864/… –  JavaRocky Apr 23 '10 at 0:36
So you won't be drilling a hole through the earth, eh? I had a perfectly good answer for you if you were but since you aren't, sorry... ;) –  joeqwerty Apr 23 '10 at 0:42
What are the chances that a giant squid gets tangled in this cable? That's probably going to throw off your calculations. –  joeqwerty Apr 23 '10 at 0:45
@joequerty Lets also consider drilling has become an illegal activity. "There is no drill! [bathroom]" youtube.com/watch?v=WjXzXUuyAgI –  JavaRocky Apr 23 '10 at 1:05
@joequerty I ate all the squid with salt and pepper. Sorry. Let's just consider the scenario :D –  JavaRocky Apr 23 '10 at 1:05

4 Answers 4

up vote 4 down vote accepted

Actually, it does not even travel close to the speed of light.

Firstly, the refractive index for glass is approximately 1.5, varying by the exact material. This means that light traveling through glass in a straight line only goes about 2e8 m/s, not 3e8.

And then you have to consider the curvature of the path. Even with single-mode fiber there is a lot of sideways drift along the entire run. This can easily increase the actual distance traveled by 15% or more. So what sounds like an incredible (although it actually isn't given such large distances) 3e8 m/s has been cut nearly in half.

Transmission with lasers and mirrors would be faster, but then you have to consider that things such as atmospheric perturbation, and even birds would cause communications trouble.

share|improve this answer
Interesting. What is sideways drift? My noobie google search terms didn't help me. –  JavaRocky Apr 23 '10 at 1:03
If you had a perfectly straight run of fiber, it would be possible for a light beam to enter at one end of the run, and exit at the other end. However, fiber is never actually straight, and because light can only go in straight lines, then you'll find that the beam goes in, runs along until it hits a side, which then cause it to bounce, and it will go off further along the fiber. –  gorilla Apr 23 '10 at 1:08
Light always travels in a straight line, unless it meets a medium interface where there exists a difference between the refractive index of the two media. If the angle is obtuse enough, the difference is large enough, and the current medium's index higher than the other then what happens is the light reflects back into the first medium. This is how a fiber optic cable works. As the cable curves, the light ray bounces side to side along the cable as it travels down it, making the distance the light ray travels longer than the length of the actual cable. –  Ignacio Vazquez-Abrams Apr 23 '10 at 1:09
Re: Sideways drift. Mind = Blown. –  JavaRocky Apr 23 '10 at 4:05
@Ignacio. Light will also curve due to gravity, not just medium refraction... en.wikipedia.org/wiki/Gravitational_lens Obviously, that has nothing to do with OP's question, but it does affect your "always" statement –  Scott Lundberg Apr 23 '10 at 19:29

Yes, definitely. The only thing you can to today is to reduce the latency added while sending or receiving the signal (faster computers, faster routers etc.).

share|improve this answer
Lets consider there are no routers between this link :) –  JavaRocky Apr 23 '10 at 1:01
Glass with lower refractive index. –  dbasnett Apr 23 '10 at 1:23
Of course, if the refractive index gets too low, the light will leak out of the cable... I'm not sure what the practical limit is. –  David Z Apr 23 '10 at 3:27
Light travels through the fiber because of internal reflection off of the cladding. –  dbasnett Apr 23 '10 at 11:03

Yes The answer is simple. Figure the speed allowing for the fact that the medium is not in a vaccuum and then calculate the distance to determine the time a trip will take. It cannot be changed.

Naturally, there are other factors for data transmission as TCP/IP is somewhat chatty

share|improve this answer

I think fiber optic technology might be improved to where it closer approaches the speed of light.

Guessing whether there might ever be a technology that can transfer data FASTER then light is purely speculation. I've heard very theoretical talk of transferring data via quantum entanglement...where two photons can be "entangled" as one object and changing photon A's state will cause photon B's state to change, even when they are seperated over a great distance. Is this faster then light? Will it ever be practical? I don't know, just my non-physicist-layman's 2 cents (see comments).

share|improve this answer
Ummm, wasn't that just some gumph made up for Mass Effect 2 to fill a neat plot hole to explain how communications were capable in real time at opposite ends of the galaxy? –  Mark Henderson Apr 23 '10 at 2:10
Actually it's real - this is basically what quantum teleportation is. But in order to transmit information via quantum entanglement you also have to send some (smaller) amount of information by a classical method, so the overall information transfer is still limited by the vacuum speed of light. (Also: @Cory J, quantum mechanics doesn't allow you to change photon A's state in a predictable manner. All you can do is measure it and see what you get, but that will be correlated with photon B's state.) –  David Z Apr 23 '10 at 3:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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