Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

SpaceX's newest Starmind will make earth data centers obsolete

teslarati.com

Post reply on HN
Print view View original

11–20 of 25 posts

Re: SpaceX's newest Starmind will make earth data centers obsolete

#11

Post by youngtaff »

pfdietz wrote: post #9
jfengel wrote: post #8

I just don't get it. It's really cool that SpaceX has the ability to put stuff in orbit without it being a billion-dollar event. But I don't see any advantage to having unmaintainable tech in orbit, where all you can do is throw away an entire satellite. What's the advantage? You can get solar power here on earth. We're not running out of land; we could put them in the middle of nowhere if we wanted, for way less tha…

Well, I don't know if they're doing this, but PV can be much lighter in space. Much of the mass (and cost) of PV on Earth is structure to support it against gravity and weather. In space, a thin film PV array could be as little as a few microns thick (using for example CdTe, which absorbs light much more strongly than silicon). Combine that with the 5-10x higher production from being in constant unfiltered sunlight,…

How are you going to get rid on the heat?

And they won't all be in constant unfiltered sunlight some will be in the earth's shadow

Re: SpaceX's newest Starmind will make earth data centers obsolete

#12

Post by pfdietz »

youngtaff wrote: post #11
pfdietz wrote: post #9

Earlier quoted context omitted.

Well, I don't know if they're doing this, but PV can be much lighter in space. Much of the mass (and cost) of PV on Earth is structure to support it against gravity and weather. In space, a thin film PV array could be as little as a few microns thick (using for example CdTe, which absorbs light much more strongly than silicon). Combine that with the 5-10x higher production from being in constant unfiltered sunlight,…

How are you going to get rid on the heat? And they won't all be in constant unfiltered sunlight some will be in the earth's shadow

By radiation. You know, like every other satellite does.

They will not be, or need not be, in orbits with substantial amounts of shadowed time.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#13

Post by pfdietz »

youngtaff wrote: post #10
pfdietz wrote: post #6

Earlier quoted context omitted.

Why is 20,000 launches a load of nonsense?

20,000 launches is 11 Starship launches a day for 5 years It's an unsustainable rate and even if we do get to that rate it'll wreck the planet

Why is that unsustainable? You need to present an argument that is more than just "large number! wave hands!"

It's just like the arguments used by anti-renewable energy forces here on Earth. "Large number (of PV modules and wind turbines and batteries)! Unpossible!"

And no, that launch rate does not "wreck the planet".

Re: SpaceX's newest Starmind will make earth data centers obsolete

#14

Post by jfengel »

pfdietz wrote: post #12
youngtaff wrote: post #11

Earlier quoted context omitted.

How are you going to get rid on the heat? And they won't all be in constant unfiltered sunlight some will be in the earth's shadow

By radiation. You know, like every other satellite does. They will not be, or need not be, in orbits with substantial amounts of shadowed time.

Most satellites aren't intended to produce thousands of watts of heat.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#15

Post by pfdietz »

jfengel wrote: post #14
pfdietz wrote: post #12

Earlier quoted context omitted.

By radiation. You know, like every other satellite does. They will not be, or need not be, in orbits with substantial amounts of shadowed time.

Most satellites aren't intended to produce thousands of watts of heat.

So now we're quibbling over numbers, not saying the problem is impossible.

These satellite will need larger radiators. It's an issue satellite designers have long dealt with, so the presumption that SpaceX has somehow missed the issue in their new design is not credible.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#16

Post by deeg »

pfdietz wrote: post #15
jfengel wrote: post #14

Earlier quoted context omitted.

Most satellites aren't intended to produce thousands of watts of heat.

So now we're quibbling over numbers, not saying the problem is impossible. These satellite will need larger radiators. It's an issue satellite designers have long dealt with, so the presumption that SpaceX has somehow missed the issue in their new design is not credible.

If I had any trust in Musk-run companies then I would agree with you but I think its closer to the hyperloop than reality.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#17

Post by dlcarrier »

jfengel wrote: post #8

I just don't get it. It's really cool that SpaceX has the ability to put stuff in orbit without it being a billion-dollar event. But I don't see any advantage to having unmaintainable tech in orbit, where all you can do is throw away an entire satellite. What's the advantage? You can get solar power here on earth. We're not running out of land; we could put them in the middle of nowhere if we wanted, for way less tha…

The advantage is that it's not in anyone's backyard. No where in the US is in-the-middle-of-nowhere enough for developers to avoid backlash. If it weren't for that, orbital data centers would not be cost effective. In the real world, they likely are, especially as the cost of launching satellites has plummeted.

It's the same boat as nuclear power plants. There's no economical reason to build small modular independent nuclear reactors. It makes much more financial sense to combine as much infrastructure as possible, even when the reaction chambers themselves are small and modular, like a CANDU reactor. Yet, getting approval for a few small things is much easier than one big thing, and getting approval is what's preventing nuclear power from being cost effective, so the most cost effective option is to spend more on infrastructure, so you can much, much less getting approval.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#18

Post by dlcarrier »

avgarrison wrote: post #2

How are they going to deal with the heat dissipation problem?

Put a radiator in the shadow of the satellite. The radiator size needed is similar to the solar panel size. Look at the International Space Station; the solar panels are larger than the radiators, and the occupants are far more heat sensitive than semiconductors. The hotter the radiator, the better it radiates.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#19

Post by dlcarrier »

chancitag wrote: post #5
sfmz wrote: post #4

Earlier quoted context omitted.

They already have space-based compute in terms of Starlink and Starshield. Is it much more difficult than that ?

It is much more difficult than that. Starlink is essentially IT infrastructure, which certainly produces heat but nothing on the level of pure compute. Ejecting heat in space is a difficult problem that is currently solved on the ISS with large IR radiators which take up weight and space. The size, weight, power, cost tradeoffs lean heavily in favor of ground-based compute.

The ISS dedicates more space to solar panels than radiators, despite the occupants having a very tight temperature requirement, which is about 20% cooler than the max most semiconductors can indefinitely run at. Just putting the radiator on the surface of the satellite might be enough, without even needing a panel.

Even if a heat radiator as large as a solar panel is needed, it'd be only a small portion of the overall cost.

Radiating waste heat on earth is a much more substantial portion of the total cost, especially so considering the social backlash it creates.

Re: SpaceX's newest Starmind will make earth data centers obsolete

#20

Post by dlcarrier »

youngtaff wrote: post #3

According to the authors numbers it'll take at least 20,000 launched to put a million satellite's in orbit And that's before you consider the issues of operating a DC in space It's just a a load of nonsense

With a max of one million satellites in the constellation, if Space X targets the same 5 year lifespan of its current satellites, that would be 4,000 launches a year.

In 2020, Space X had 25 launches, then 31 in 2021, then 61, then 96, then 136, and last year launched 167. That's an average of around 50% more per year. At that rate, to get to the 24x increase needed to launch faster than the constellation can support, they'll only need to keep on that trajectory for eight more years.

Post reply on HN