Hacker News new | ask | show | jobs
by steckerbrett 3795 days ago
> 50% OF US LIVE NEAR THE COAST. WHY DOESN'T OUR DATA?

It's corrosive, expensive to get things to and from it for replacement, leaks destroy the hardware, it's not close to power generation, internet access needs cables because RF doesn't penetrate water, everything is going need watercooling which is rather expensive.

11 comments

Imagining they completely solve the problem of sea water, leaks, etc, it still is amazing to think that you would want to do your server maint by pulling a data center out of the ocean on a boat and replacing hard drives and the like.

The only way this makes sense to me is if there is the ability to create something akin to the cargo container as a building block of a data center, where you can have arbitrary compute and storage plug into a greater complex.

Sounds like that's exactly what they want to do: they would only pull them out of the water every 5 years to do computer replacements / maintenance. If some components fail then who cares. They wouldn't do a full rebuild for 20 years.
I worked in large data centers before and I just don't see how this can be done practically. Data centers require quite a bit of physical maintenance.

Every computer design has some element that will render a large part of the design inoperable in case of failure. Either it is a SAN head (even if you have two, the fail over can malfunction), or a switch setup.

Then there are things like failures of simultaneously purchased components (hard drives purchased at the same time, that are worked the same load will roughly fail at the time).

Cloud datacenters are not complex heterogeneous mixes of components. There's no SAN head. It's one thing multiplied + some networking gear. Even if a top of rack switch fails they're still not going to yank the box yet because the TCO will be lowered by too much maintenance at this scale. They wait for their maint interval and fix everything at once (or just upgrade the hardware).
Think of a farm of small data center pods with cloud apps. When failure in a pod exceed useful threshold, apps are migrated out to other pods and the pod is retrieved, serviced and returned to its place.

A custom made barge with dynamic positioning gear and a grabbing/coupling system to detach the pod from the subsea grid, lift it, and then re-attach it would make the servicing relatively efficient.

I could see the roundtrip time for a full hardware replacement of a pod being under an hour, conceivably under 10-15 minutes.

> The only way this makes sense to me is if there is the ability to create something akin to the cargo container as a building block of a data center

Which is something Google already did[1].

1: https://en.wikipedia.org/wiki/Google_Modular_Data_Center

And if it is just cooling, why aren't datacenters built on the coast pumping sea water for cooling?
High cost of the land on the coasts?
And I guess you wouldn't even need to use sea water as a primary cooler, just as a secondary cooler. I.e. the primary cooler flows through your datacenter and the secondary cooler cools that primary cooler. So fewer pipes are exposed to sea salt.
So, same procedure as power plants.
Yeah. Instead of trying to protect the environment from what's inside (radioactivity), you are trying to protect what's inside from the environment (sea salt)!
I think they should be able to handle most of those issues, for example they may be able to use wave power for power generation. Furthermore, I don't see how RF opacity is an issue, seeing as anyone running a data center over RF is criminally insane.

edit: I can't grammar

You run management connections from separate computers over RF as a backup and for fail-over in event everything else fails. I mentioned that in a recent comment related to Github outage and preventing those:

https://news.ycombinator.com/item?id=10996442

Can help with certain security situations, too. Not sure how much that applies if it's underwater, though. Most infiltrations would probably turn into a denial-of-service attack effectively haha.

Fair enough. It doesn't get much more literally out of band, eh?
I tried to get it further by proposing a neutrino-based communication system. Can just send the signals straight through the planet itself to a datacenter very far away. I was told there would be both implementation problems and cost overruns with that project. Went back to default recommendations for wireless.
> everything is going need watercooling which is rather expensive

Why would everything need water cooling? I'd expect that something using water would be used to keep the air inside the unit cool, and then the cooling for the servers themselves would be ordinary air cooling.

Assuming they have ways around some of those issues this could work out rather well. The important thing is that it's only a research project. Microsoft's research turns out some really awesome stuff but plenty of it failed or is cut. Who knows what'll happen to this but it's a really interesting proposition!
For power you could use a small nuclear reactor, just like submarines and aircraft carriers. I've always thought it would be a fun exercise to take a decommissioned nuclear submarine and turn it into a floating datacenter.
When nuclear vessels are decommissioned they remove the reactors. Operating a naval reactor requires a constant watch by multiple highly-paid experts. They are not cost effective for electrical power generation.
The staffing required to operate a nuclear submarine is astronomical, so it would set the person:server node ratio back decades.

It might be fun but there are better things you can do with $700M.

Yeah but we only need the reactor, not the whole submarine.
Do you want the raft? Because that's how you get the raft.
Because 50% of us live near the coast, not past it. Maintaining anything in close association with an ocean is painful. Everything rusts. Even the stuff they say doesn't does. Anything that moves ages at an accelerated rate. As soon as the slightest waves start, little salt crystals appear on every surface.
> everything is going need watercooling which is rather expensive

Water cooling significantly reduces running costs which is why many DCs are switching to it. Over the long term you save money.

I've read that ancient Roman concrete was manufactured in a way that is seemingly lost to time. It's also practically impervious to the elements.

http://www.romanconcrete.com/docs/spillway/spillway.htm

If we want to find a way to build beneath the sea that's a good place to start.

According to this article, a 2013 study successfully reverse-engineered the recipe for Roman concrete:

http://www.ancient-origins.net/news-history-archaeology/rese...

That is an awesome link. Thanks for that.
Do data centers ever rely on RF for connectivity?
No, but it's a harder proposition to have fibre runs to a server which is in the ocean. You can throw a normal server somewhere silly and connect to it wirelessly, this is just the one remarkable exception.
What if you hook up to some of the undersea fiber that's already there? (I still think this is a bad idea)
I don't think that's possible, it would be sheathed in extremely thick steel and not something you can just splice onto.
The NSA disagrees.
Same question in my head, how do they prevent corrosition?

And why "everything is going need watercooling which is rather expensive" ?? If i remember correctly every OVH DC Server are using it.

For marine gear you either use a material which won't readily corrode, or you use sacrificial anodes which are galvanically consumed rather than something you care about.
Seriously. It's a proposition obviously cooked up by someone who has never spent more than a week's vacation by the sea.
-> It all started in 2013 when Microsoft employee, Sean James, who served on a US Navy submarine submitted a ThinkWeek Paper.
Well Seattle is right by the sea, so..
Microsoft isn't in Seattle, so...
I'm aware - I worked there. But it's pretty close - certainly not far enough to be oblivious to the ocean
It's a 20 minute drive from Redmond to Seattle.

It's a 3 hour drive from Seattle to the ocean.

I suppose it would depend on what you'd qualify as "the ocean". I lived in Seattle for 4.5 years, and while I don't consider Elliot Bay[1] to be "the ocean" per se, it's pretty close - and you have all the corrosion problems and general exposure to the elements that you'd get, which I think was the original point of the comparison in this thread.

[1] https://en.wikipedia.org/wiki/Elliott_Bay

Ok someone has never been to Seattle it seems. It's not a 3 hour drive to the ocean.
It's amazing that the drive to the Northern part of the Olympia Peninsula is so long, Bellingham which looks much farther is around a half hour closer.
With minimal traffic, it is: https://tinyurl.com/zbkp89q