[erid]

The latency must be huge though.

[paulgerr]

2.6 second round trip for light, so pretty huge ping times.

[nraynaud]

That's why I always found moon people to be lame at quake!

[Aardwolf]

One day technology will really have to do something about lightspeed to allow interplanetary quake!

[deletes]

Solution for the lag will be a special category "bot play". The bot will not be programed but will behave like a black box, computer ai will record you every time you play in a local game and will slowly build the bot around. Then the bots will compete on a server, maybe separated by categories determined by previous recording time.

[rel]

I worked at NASA (AMES) on a project to put small mobile robots on the moon. Our biggest goal was to create a tele-operated mode where from earth we could actively roam around the moon. The hardest part was that factoring in a 4 second delay from our commands to the network stacks to the rover on the moon, and then that video from the moon back to our computers and then creating good judgement. A 4 second delay seems insignificant but is in fact detrimental to intuitive controlling.

[rodh]

I'd imagine TCP, as a protocol starts to fall over with moon-to-earth latencies.

Or would it? Perhaps simply increasing the retransmission timer would give us interplanetary internet. I wonder.

[drcross]

http://en.wikipedia.org/wiki/Interplanetary_Internet

[noselasd]

Code from NASA here, http://sourceforge.net/projects/ion-dtn/

[TylerE]

Moon to earth should be workable enough...average speed of light roundtrip will be about 2400ms.

[throwaway9101]

Bandwidth delay product tends to limit throughput at such high latencies, though.

[stonemetal]

TCP uses RTT time to try to guess network congestion. Having worked on software that almost always had satellite hops, it sucks on long links. Primarily what you notice is that you get limited bandwidth because the TCP stack is throttling to prevent congestion. Not sure what RTT is to the moon and back but you would want to use something like TCP Hybla that does network congestion control a different way.

http://en.wikipedia.org/wiki/TCP_congestion-avoidance_algori...

[arnarbi]

> interplanetary internet

Winternet!

[7952]

The latency is sufficient that you could actually use space as a data storage medium. At 622Mbps there is 1,596Mb in transit at any given time.

[gprasanth]

Unless, our technology advances to always do stuff a couple of seconds in advance by knowing our intents.

[toomuchtodo]

Or we figure out how to use quantum entanglement or gravity as information transfer mechanisms.

[danbruc]

Gravity propagates at the speed of light, too.

[swamp40]

I think that's a debatable topic.

Some say it is just a pure geometric effect of curved space (and thus instantaneous), not a force of nature that propagates.

[danbruc]

http://en.wikipedia.org/wiki/Speed_of_gravity

[wiml]

The pure geometric effect propagates at light speed also (unless you want to discard relativity).

[chrischen]

Then we can tug on gravity to send signals instantaneously!

[drakeandrews]

Unless we discover gravity is bound to the speed of light too.

[stonemetal]

http://www.newscientist.com/article/dn3232-first-speed-of-gr...

It is.

[sslayer]

Spooky TCP

[toomuchtodo]

Someone get Vint Cerf on the line; this will make a fine RFC.

[rzimmerman]

Still not as bad as Voyager, where the ping time is an entire weekend.

[agashka]

wouldn't it be less then RF? Radio Frequencies travel at the speed of sound, while lasers would be light-speed, from my understanding. I'd really love some insights on this!

EDIT: I still need to learn a lot on signals!

[zargon]

Radio is not sound, it is an electromagnetic wave that travels at the speed of light.

[robin_reala]

Radio waves are EM waves, same as light. They all travel at c.

[16s]

Speed of light, not sound.

[robodale]

oh boy.