By way of another question, teleporter stations exist at locations separated by tens of light years.

But if it takes ten years to reach your destination then that's ten years in which the receiving station might have been put out of commission, plus another ten years to find out, so a 20 year leap of faith.

If somebody steps into a teleporter and beams themself to a destination ten light years away, their original self is destroyed (nevermind the philosophy, they made their choice) in anticipation of it being successfully reproduced at the receiving end. If the receiving end is switched off then they perish (or technically continue their journey into uncharted space until the signal his dispersed irrecoverably).

I assume that the teleporter maintains more-or-less continuous operation, sending and receiving things day after day with no way of being certain that things it sends are arriving; but it can't very well keep the original and continue to call itself a teleporter (that's a separate question), and it's way too much data to record for local reconstruction in cask of a NACK return.

I propose, as a safety measure, a string of relay stations between major destinations. Each relay in continuous communications with its neighbours to confirm that they're still operational. If such a relay receives a person (or whatever) from one neighbour but realises that the other neighbour is down, it diverts that person back where they came from, because that's preferable to killing them or stranding them on the network light-years from anywhere.

The transit times are already terrible, so we don't want to make this worse by more than [arbitrary figure alert] 0.1%. If each node were to add a delay of six seconds then they must be at least 100 light-minutes apart (5k nodes per light-year), but probably much further because they're costly. On the other hand, the cost of repair -- should it come up often -- might be dominated by the conventional travel time from the last working node to the broken node if they're too sparse. Also the window of irrecoverable loss is larger.

What are the considerations for maintaining such a network once it's established? What's going to interfere with it, and how might that be corrected?

  • 3
    $\begingroup$ You might want to study how current internet is done, including all ISO-OSI layers $\endgroup$ Jan 28 '16 at 8:31
  • 1
    $\begingroup$ What @PavelJanicek said and also look at how we are considering doing space internet: en.wikipedia.org/wiki/Interplanetary_Internet $\endgroup$
    – Selenog
    Jan 28 '16 at 11:35
  • $\begingroup$ A major factor in this network would be what the recovery process looks like on a failure, when you send someone to a station that doesn't exist. You still have a large number of smaller hops, each of which has hardware that can fail. In general, we design our networks to combat the actual failure mechanisms seen, so we need to know what the failure mechanisms are, and how bad they are. $\endgroup$
    – Cort Ammon
    Jan 28 '16 at 18:35
  • $\begingroup$ @PavelJanicek Terrestrial internet (or any plain-data network) didn't seem to offer much insight, as it rarely has to make an effort to salvage something that was sent out years ago. Think of all the teleportations that might have been made in the 20 years before you realise the far end is down; and you want as much as possible of that to be rescued. $\endgroup$
    – sh1
    Jan 28 '16 at 18:49
  • $\begingroup$ @CortAmmon That's the sort of consideration I'm asking about. What sort of failure modes might one anticipate in interstellar space? What has to be accounted for. Is a node likely to be destroyed or moved by anything passing by? Does anything ever pass by? Are stars moving too quickly to track, and does the chain of nodes improve that situation? I'm guessing fuel (nuclear?) has to be teleported out to all the nodes periodically, but with the supply line being the communication line, could an outage starve the rest of the chain to death? etc.. $\endgroup$
    – sh1
    Jan 28 '16 at 18:55

Avoiding mechanical failure
I propose a system of switchable receivers. This answer is built on one assumption: If we send a package between teleporter A and B, and teleporter B is broken, the package isn't instantly lost. Instead it will only be lost when it tries to arrive at a broken teleporter. If B is fixed while the package is in transit, everything should be fine.

A teleport station should not consist of a single teleporter, instead it should consist of, say, five. When you want to send a package between station A and station B, you first send five test packages to all five of B's teleporters. Then you wait an hour and send your important package to teleporter B-1. At station B, they will receive the test packages indicating that an important package is coming.

If, say teleporters one two and four have failed, you'll observe that you only got a test package in teleporters three and five. You then have an hour to swap out teleporter three with the failed teleporter one, so that there will be a working teleporter in place to receive the incoming important package.

This system ensures that mechanical failure will only cause lost packages if all five teleporters suffer a simultaneous failure, which is astronomically unlikely. If you're concerned about the buildings housing these teleporters failing, you can put each teleporter in a different building (or heck, in space stations) and arrange for a system to automatically send a working teleporter to intercept the package headed for a broken one.

This also means mechanical failure will be easy to fix: If one teleporter fails, it's still part of a station with other working teleporters, so the mechanic can use one of those to arrive and make repairs.


Alright, so we're protected against random mechanical failures. But what if we don't trust the people on the other side of the teleporter? What if we're afraid they're going to shut down the teleporter and destroy all incoming cargo? Well there's no way to prevent that happening, but we can build a relay network to minimize the amount of cargo traveling to an untrusted teleporter, meaning we'll lose less if those jerks on the other side sabotage us. Once our network learns that the other end has been turned off, it can simply return packages to sender.

The obvious problem is that if we don't trust the other side, they probably don't trust us, but someone has to control the relay network. No matter how it's split, someone has to control at least 50% of it, which means that someone could decide to turn off all their teleporters and destroy 50% of in-transit packages, no matter how dense the relay network is.

The solution to that is to build two networks. Planet Zebrox has a network for sending things to Earth, which they control and trust, and Earth has a network for sending things to Zebrox, which Earth controls and trusts.

How to make a relay network, cheaply

The problem is that we'd like to have a low volume of untrusted traffic (say, one hour's worth), but we don't want to build a relay node every light-hour along the way between Earth and Zebrox. Luckily there's a solution.

Let's say planet Zebrox is ten lighyears away. Here's how we build Earth's relay network.

  • Teleport Station A on the surface Earth, sending packages to B
  • Teleport Station B, 1 lightyear away from Zebrox, linking A and C
  • Teleport Station C, 1 lightmonth away from Zebrox, linking C and D
  • Teleport Station D, 3 lightdays away from Zebrox, linking D and E
  • Teleport Station E, 7 lighthours away from Zebrox, linking E and F
  • Teleport Station F, 40 lightminutes away from Zebrox, linking F and G
  • Teleport Station G on the surface of Zebrox

Each of these stations is equipped with powerful telescopes to detect any ships approaching it, and it also broadcasts a "Please don't come near me" signal to avoid people accidentally approaching it. A key component of this strategy is that space travel is slow, there are no ships buzzing around at half the speed of light.

Let's say that, simplistically, the maximum speed a ship can reach is 0.04 c. If Station C that sees no ships within 1 lightmonth of itself can be confident that it will remain safe for the next two years or so. It then sends a signal telling that to Station B. Station B receives the message a year after it is sent, and knows that if B sent a package to C now, no one could possibly reach C and sabotage it before the package arrives, so it can safely send to C.

Similarly, D sees no nearby ships and tells C that it's safe to send a package because no one could possible seize control of the station before the package arrives. This trend continues for E and F.

In this way, we can ensure a trust relationship that lasts for almost the entire length of the journey, meaning we can only ever lose the small percentage of traffic flowing between F and G (plus the additional amount of traffic F would send to G before hearing that the Zebroxians sabotaged G).

It might be a bit implausible for station B to be able to detect all ships within a lightyear of it, the limitations of telescopes being what they are, but if we say that telescopes can only detect ships within one lightmonth of them, then we can set up the following array:

  • Teleport Station A on the surface Earth, sending packages to B
  • Teleport Station B, 9 lightyears away from Zebrox, linking A and C
  • Teleport Station C, 8 lightyears away from Zebrox, linking B and D
  • ...
  • Teleport Station J, 1 lightyear away from Zebrox, linking I and K
  • Teleport Station K, 1 lightmonth away from Zebrox, linking J and L
  • Teleport Station L, 3 lightdays away from Zebrox, linking K and M
  • Teleport Station M, 7 lighthours away from Zebrox, linking L and N
  • Teleport Station N, 40 lightminutes away from Zebrox, linking M and O
  • Teleport Station O on the surface of Zebrox

This still ensures the trust relationship is only broken for a distance of 40 lightminutes, without the incredible expense of building a relay node every 40 lightminutes.

  • $\begingroup$ One of the failure modes that worried me was the risk of the far-end recipient suffering some kind of revolution (as per the seed question, it may be hard to manage) and deliberately turning the machines off. You won't know it's happened for 20 years, and if you have a busy transit/supply cycle running to the destination then it's 20 years of people and goods that you'll lose if they do this. $\endgroup$
    – sh1
    Jan 28 '16 at 19:03
  • $\begingroup$ @sh1 Just how fast is non-teleport space travel? If it's slow (at most a small fraction of the speed of light), I can come up with a much more efficient relay network than "One node every X lightminutes". $\endgroup$ Jan 28 '16 at 20:43
  • $\begingroup$ I hadn't considered it. Let's say that teleportation became a thing relatively early and stalled the development of conventional human space travel. The delivery of materials and durable machinery can be sent much faster than humans but still not a substantial fraction of light-speed. $\endgroup$
    – sh1
    Jan 28 '16 at 21:32
  • $\begingroup$ @sh1 Edited with a solution to the "What if they turn off their receiver?" problem. $\endgroup$ Jan 28 '16 at 23:32

Are teleporters duplicators? Say I teleport from A to B and station C is in the middle can C generate a copy of me without disrupting the message? Can you teleport without destroying the person? They you can keep a copy in stasis until an acknowledgment of safe arrival comes.

  • $\begingroup$ If the teleportation device is in any way technological, then yes, it is a duplicator by default. Although it may not necessarily have the power to recreate all the mass, it would be capable of duplicating the data. Downside of waiting for a reply: say for example that the dest. is 10ly out. Copy arrives safely, but a link goes down and the acknowledgement doesn't get through, then a second copy is transmitted 20 years after the first in a retry attempt. $\endgroup$ Jan 28 '16 at 19:59
  • $\begingroup$ @Draco18s, worst case, if you don't have the room to keep twenty years of travellers in stasis so you just let them walk out the door wearing a beacon, then the poor old original will be walking about minding its own business back on Earth when the confirmation comes in and ZAP! Transfer compete. $\endgroup$
    – sh1
    Jan 28 '16 at 21:19

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .