I am curious if we can achieve 1 ms ping worldwide for the internet. I made a simple and not-well-informed calculation.
I assumed the speed of information to be the same as the speed of light in the fiber cables because we will use fiber cables and I don’t think there is a better technology. The speed of light in fiber cables:
So, instead of 204,190,477 m/s, for the easiness of calculation, I rounded down to 200,000,000 meter per second (two hundred million per second).
The longest line that is possible between two points on the earth should be the same as the half of the earth’s perimeter. The circumference of the earth is 40,075 km according to a quick Google search. Divide by two and round it down: 20,000 km (or 20,000,000 meter).
Divide 20,000,000 by the speed of light in the fiber cables: 0,1 second. This is 100 milliseconds (i.e. 100 ms ping). Various equipment used by ISPs adds latency, but let’s imagine we have a “groundbreaking” equipment which adds latency that is small enough to ignore.
So, is this (100 ms) the lowest number we can achieve due to the limits of physics?
Is it possible to do it better?
Are you asking about limits due to physics, or limits due to engineering? The engineering limit is signals transmitted along the surface of the Earth at the speed of light in fiber optic cables. The physics limit is signals transmitted in a straight line at the speed of light in vacuum. Which is only about a factor of two better, anyway.
Radio waves travel in vacuum at c. So a constellation of satellites in low earth orbit (like Starlink) could relay signals to the other side of the Earth faster than fiber optic cables could. The signal has to travel up to the first satellite, then take a slightly longer path around the Earth, then travel down at the end, but that only adds about 1500 km to the travel distance if the satellites were aligned perfectly. (about 300 km up, then 300*pi longer path around the Earth, then 300km down).
Of course each satellite in the link adds latency, but I think the same is true for fiber optic cables as well - it’s not a single cable going all the way around the world.
According to this, “The majority of the latency introduced by optical transmission systems is in the form of dispersion compensating fiber (DCF),” So the speed of light calculation for the main fiber is not very close to the final answer. Using optoelectronic repeaters would be even slower.
AFAIK both optical an coax have practical limit of about 2/3 of speed of light at best. You can see or hear the delay IRL. In a range of about 10ms in most cases involving half planet distances. Examples. Calling your Aunt to Australia or playing some mass computer game over some exotic server (that is why they tend to put up at least one per continent). And by my knowledge that is serious hardwired Einsteinian physical limit.
As I recall, when actual satellite systems started to replace cable (coax) systems for international communications, the satellite systems had much greater latency. I never thought about the relative wave speed: we just assumed it was the greater path length that made the latency worse.
There are two questions here , which can have different answers. If you are bounding your question by assuming you need to allow a ping to travel from the two furthest points in the world then you have all the limitations as already discussed.
But if you are asking the more practical question ‘what is the lowest ping for anyone to practically access information on the internet’ then you can use network architecture to significantly reduce that time.
Just mirror the entire internet at several different locations on the earth and you are automatically closer to everyone.
Major service providers already do something close to this with their own, massive, amounts of data ie Netflix, YouTube, etc. They are called Content Delivery Networks (CDN) https://www.cloudflare.com/learning/cdn/what-is-a-cdn/
Mirrors the entire internet would obviously be a staggering step up from this, but imo if you limited your scope to 90% of all data and put some mega billions into dedicated infrastructure, it’d be possible with today’s tech.
It’d also probably be far cheaper than the massive amount of additional fibre and/or microwave links needed for any meaningful worldwide ping reduction.
That’s because most telecommunications satellites are in geostationary orbit, 35,786 km above the equator. I’m talking about comsats in low earth orbit, which is a fairly recent development because you need a huge constellation of satellites to have good coverage.
The stuff for which you care most about ping times is stuff which inherently can’t be mirrored. If Netflix or YouTube has a 1000-ms ping, then that just means that my 30 minute video starts playing a second later than it would have. On the other hand, if I’m playing a competitive video game with a 1000-ms ping, then I’m dead before I ever even saw the other guy.
I am curious about the microwave (and other RF) links. For long distances, you need multiple repeaters. Do these repeaters store-and-forward data, or re-transmit it without delay? If they achieve one-bit-in, one-bit out near-simultaneously, this would be a far faster throughput than if they receive a packet, then re-transmit that packet, even without any packet analysis delay.
The spooky quantum stuff is useless here. Quantum entanglement does potentially have its uses in communication, but mostly for safeguarding information against eavesdropping. You can’t use it to transmit useful information faster than light.
True - but microwave transmission speed is still significantly faster than fiber. And with low-orbit satellite constellations getting in on the act, the actual number of hops can be reduced, and so capitalizing on the higher link speed.
And a bit of googling suggests that high performance microwave repeaters are capable of operating in the “low nanoseconds” of per-hop latency.