If I asked the same question for operating systems, I'd say I am quite confident that some descendants of Linux could be there for a very long time.

Today, while Linux is used in embedded IoT devices, it is not yet used to operate the ISS. Also, a complete redevelopment of the Linux kernel is estimated to cost a considerable amount of money.

But then if I think of first colonies and their IT infrastructure, would they need some sort of IaaS (infrastructure as a service) and virtual machines?

I think they would, due to reusability and reproducibility and high automation capabilities. But I think they're not there yet with requirements for reliability for living in space.

But then, if so, what should an outer space data center look like? Will it be a commercial black box facility delivered by one of the usual suspects and, if not, could we speculate whether it would be something completely different, or a descendant of today's open source cloud software? And how should we recognize it?


closed as too broad by Mołot, Gryphon, Frostfyre, Agrajag, GerardFalla Feb 11 at 17:40

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ Hi J. Doe, I made an edit to your post for grammar and clarity. If I got anything wrong, please change it back. I see we made the same change to the title at the same time, but it looks like both edits went through fine. $\endgroup$ – Cyn Feb 11 at 16:01
  • 3
    $\begingroup$ I think that if you narrowed down your scope this question would be a better fit for this SE. I am afraid that, as it is now, it asks for too many problems and mostly based on assumptions you don't specify. $\endgroup$ – L.Dutch Feb 11 at 16:05
  • 1
    $\begingroup$ As a point of reference, we tend to use complexity to make more complexity - which is something the space program generally can't afford. To make the point: the HP-41CV calculator I had in high school had more computational power than any of the Apollo space capsules, and yet those capsules did their job just dandy. $\endgroup$ – JBH Feb 11 at 18:26

Virtualization would be a part of space colonization if virtualization would be part of space colonization. Sadly, that tautology is really the best we can do.

Virtualization is a tool which can be used to solve a problem. If we have that problem, then we can use that tool. If we don't have that problem, then we wont use the tool. If there are additional requirements which prevent use of the tool, then we can't use the tool.

Virtualization will be popular if we need to have the things virtualization supports. So far, that isn't something space travel has needed. The biggest reason we haven't seen it is because of complexity. Current space software designers are big fans of simplicity. Virtualization is just one more layer that could go wrong. Also, current space software is designed on an enormous budget. When you're managing satellites whose cost plus launch can be a billion dollars, you can afford to spend the time verifying the software works. Its unlikely that you will see software which is untrusted enough to benefit from virtualization.

However, if space travel becomes truly commercialized, which your talk of "colonies" suggests is the case, then you're going to see commercial grade datacenters. On that day, you will see commercial grade solutions, like virtualization.

Personally, I don't think virtualization will be part of space colonization because we're going to come up with whatever "the next big thing" is. Virtualization solves a few very thorny problems, but it's not the only kid on the block. We probably have a lot of time before space colonization. The old days of dial up modems may be closer to our present day than colonization. A lot changed since then, and a lot will change.

If I had to pick a solution from today to appear in space colonization, I'd expect it more likely to see something like Project Midori from Microsoft. Midori went in the absolute opposite direction of virtualization. Instead of creating layers which try to perfectly encapsulate untrusted software and never let it free, Midori lets all applications intermingle in one big address space, and relies on tools to prove that one application can't mess with another. Midori was based around C#, a managed language, and they used the managed language features instead of virtualization.

The result? Midori's code runs a little slower because it's managed (and you can't fine tune managed code quite as precisely as unmanged code), but because it can safely put all processes in one address space, there's no need for expensive context switches -- all your programs can effectively run at the kernel level!

As for technologies which don't exist yet? I'd say the Internet of Things is demonstrating a major security issue which isn't solved via virtualization. We don't yet know how to secure a few hundred small devices all connected together on a network. Whatever solution we come up with for that is likely to appear in space colonization as we progress.

  • $\begingroup$ To add to this, watch the OVirt mailing list to see the issues that people had updating from OVirt 4.2 to 4.3. Applying this risky an upgrade to life critical engineering controls in a space context is a non-starter IMHO. $\endgroup$ – pojo-guy Feb 11 at 17:32

Not the answer you're looking for? Browse other questions tagged or ask your own question.