I’m not sure I agree. 30-40 years isn’t THAT long of a time. Besides, the point of rebuilding a global network from scratch is to fix all the problems with the old global network (the Internet) that the original architects never envisioned at the time.
Thus, it would be foolish to not try to envision future challenges like interplanetary comms when building a new design. It’s important to look ahead and design not just for the present, but for the future, to make it sustainable.
I don’t like the idea of having to build a new global networks from scratch every 40 years Let’s make SAFE as sustainable and scalable as possible now. Why not? It’s still being built, it’s not too late for these considerations. And then we won’t need to waste time building a new one later.
That said, I’ve wondered about the very same questions @Zoki has. The problem I see with one SAFE network that spans the (let’s say, Solar System) is that as the maximum distance between two humans grows larger and larger, the network may get slower and slower on the average (assuming my following thinking is correct).
My thinking is that, since chunks are spread over XOR space, its statistically likely that they’ll be very geographically disbursed. If we’re only on Earth, that’s not so bad. But let’s have a thought experiment.
Suppose it’s the year 2217 (200 years from now). Humans live (and farm for SAFE) on Earth, Mars, Saturn’s moons, and Venus’ atmosphere (in kick-ass floating blimp cities).
Now, suppose you live on Mars. You upload a file to the SAFE Network. It’s split into 3 chunks. One chunk is stored on a farm on Mars, locally. But one chunk lands on a vault on Venus and the last one on Saturn.
It would take hours to retrieve those chunks and load the file back.
Sure, you might say. But, chunks are also replicated for redundancy. One copy of Chunk #2 might be on Venus. But the other copies might be stored even closer, and SAFE always fetches the fastest copy of each chunk.
True, but even still, suppose there are 30 billion vaults or more spread out across the whole solar system. Maybe all the copies of Chunk #2 are as far away as Venus or further. That could be possible. Maybe even likely.
If there are only 1 billion vaults on Mars, then there would be 29 billion vaults off-planet. How likely would any of your chunks be stored on your planet then?
I suppose it all boils down to the statistics of chunk distributions, which I can’t say I’m an expert on. But assuming a distribution model like this, then it’s logical that the greater humanity’s breadth is across the cosmos, the slower SAFE would become (again, on the average) as chunks get statistically further and further apart.
I would very much welcome counter-arguments, because this seems like a bleak picture to me!