This is interesting and a new application of this type of communication. The graphic really buried the lead though. I've added some notes in RED that put it into proportion:

JAXA is receiving a signal from point A on Earth in a "close satellite" that is only a few hundred kilometers above the surface of the Earth in LEO (Low Earth Orbit). The "close satellite" is then sending the data via laser away from Earth to the "far satellite" about 34,600km behind it in GEO (Geostationary Orbit). The "far satellite" is then turning around and shooting that back at a different point on Earth, point B. Pretty cool stuff!

Note, the data rate is pretty high here 1.8Gbps, but because they're shooting out to GEO and back, the latency takes a big hit. We're talking about 500ms of latency. We're facing the limitation of how slow the speed of light is here. Keep in mind, for the application here that JAXA is doing, big data dumps, latency isn't really an issue. You pay the one-time (and one-way) latency cost only once when you start the transmission of data. So if you were using this to download a large file which took 10 minutes to download at 1.8Gbps, with the latency costs its only going to take 1/4 of a second longer through this satellite link for the GEO portion of the transfer vs two systems sitting 1m away from each other on a desk on Earth somewhere.

[–] surph_ninja 2 points 1 week ago* (last edited 1 week ago) (1 children)

Is there any limit on distance? Would a single pair of nodes be hypothetically able to communicate at interstellar distances like this, or would a bridge of nodes be needed?

Let’s start building nodes out to Mars, and establish the solar system’s first interplanetary internet. Once we have it down, start shooting nodes out to the nearest stars, with instructions on how to call us. It’ll be an awkward conversation with the years of delay on getting any reply, but it might not be nothing.

And if we ever manage to build autonomous space mining/refining/manufacturing, we could potentially launch an army of bots to nearby stars to build exploration satellites in each location, receiving long delayed instructions and sending back data along a laser network. We could explore the cosmic neighborhood from the comfort of our own planet.

Not something we’d be able to enjoy the fruits of within our lifetime, but a seed worth planting for future generations.

[–] partial_accumen 3 points 1 week ago

Is there any limit on distance?

Laymen knowledge here. My guess on the limitation is either on the targeting abilities of both nodes being able to get absolutely perfectly aligned down to a mm across stellar distances for maximum amount of light emitted by the sending laser to the optical receiving satellite OR a lens on the sending laser that spreads the laser light beam out a bit so alignment doesn't have to be as precise.

it is essential to continuously direct the laser beam, which expands only about 500 m, accurately to the counterpart satellite even at a distance of 40,000 km.

The cost to the beam spreading is that there are fewer photons per square mm to impact the optical receiver. So in this case the receiving satellite only has to be within a 500m cone to receive the signal, but thats at a distance of 40,000km. 40,000km is chump change in solar system or stellar distances, so that cone of light would be dramatically larger if the satellite was farther away, which means dramatically fewer photons to receive.