3 ms·

It's not the hardware, it's the crap that runs on it. You want a system which is secure and usable? Time for the DoD to commission their own OS. "Nobody will

by dsr_ 5y ago

It's not the hardware, it's the crap that runs on it. You want a system which is secure and usable? Time for the DoD to commission their own OS. "Nobody will know how to use it!" is an objection which can be overcome by training, which is a thing that the armed forces understand. "It won't be compatible with the COTS!" is an advantage, not a disadvantage. "It will take too long and cost too much!" means that they aren't in a good position now and lots of jobs will be created. "Nobody wants to work on a military system!" is technically true, but run it out of NASA or the Department of Energy and develop it with an open source license (effectively a necessity, anyway). Also, it turns out that lots of people and companies will work on things when money is involved.

snowwrestler 5y ago
OS does not matter anymore, it is a commodity. That’s why copies of any OS cost almost nothing at retail. What matters is the configuration of the OS and the network, and most importantly the selection and configuration of applications, whether hosted or client (or both). That is where DoD should focus their resources.
Jtsummers 5y ago
COTS is useful in many circumstances, it is much quicker to field if you can reuse existing commercial software. The only advantage to being incompatible with COTS is that it would feed the defense contractor industry as they'd need replacements or compatibility layers that some contractor would happily provide for $1 billion and 3 years (with an extra $100 million/year after it's late).
willcipriano 5y ago
> commission their own OS. I'd advise they start with a Linux distro. They can audit/replace components as they go to increase security over time. Even North Korea wasn't paranoid to build it's own OS[0]. [0]https://en.wikipedia.org/wiki/Red_Star_OS https://en.wikipedia.org/wiki/Red_Star_OS
jasonjayr 5y ago
Linux + wireguard-only network interfaces? With physically secured, uncloneable memory cards for each device's/interface's private key?
wyager 5y ago
> lots of jobs will be created Invoking the broken window fallacy is not a good sign for the coherence of your argument. I'm not entirely sure whether I should interpret this post literally or as a satire of NIH syndrome.
- dsr_ 5y ago
  Literal, please. Competing software ecosystems can constitute valuable economic activity. Broken Windows is, if anything, the current state of affairs: running important things on Microsoft Windows is becoming intolerable, with layers of expensive (dollars and computrons) fixes being applied. Is it economically better to start over or to keep trying to defend the indefensible?
41b696ef1113 5y ago
>Time for the DoD to commission their own OS. I am shocked that governments are not investing mega-bucks into getting a microkernel OS built that could be run on internet routers, tanks, power plants, aircraft, water treatment facilities etc. Even if microkerenel design has some impossible to overcome performance limitations, for utmost security, it would be a small price to pay knowing that hacking the OS was nearly impossible.
- d_runs_far 5y ago
  Your shock assumes there is leadership at the top that sees the big picture with respect to enabling technology. While not in the US, I have seen the inner workings of other nations armed forces, and there is an overwhelming urge to farm everything out to contractors - so there are going to be at least two competing microkernel designs (proprietary of course) for each the army, navy and air forces under the current leadership structures.
- marcosdumay 5y ago
  > Even if microkerenel design has some impossible to overcome performance limitations It doesn't. There's no inherent hit on microkernel performance. What exists are limitations of hardware that wasn't designed for it, and a complete lack of interest on creating hardware optimized for them.
V_Terranova_Jr 5y ago
My god, I cannot second this enough. We talk craptons about "innovation" and torrents of words on "cybersecurity" yet underpin daily life on recursive kludges on 1980s operating system and software technology. It's quite Kafkaesque! This isn't just a matter of "spending more on IT" - they likely already spend too much. It's just spending on the wrong things. No amount of layering crap on top will overcome the rotten, slapdash foundations of systems like windows, nor the diverging motivations between MS and and their customers. Especially customers like the U.S. Government. The amount of vendor lock-in to the shitty MS ecosystem, begetting atrocities like critical data living on godforsaken Sharepoint, is profoundly saddening. It's remarkable how everyone just uses hardware that's orders of magnitude faster than 1990s computers, yet interactions are far slower. I once tried (reluctantly) connecting my USG-issued laptop onto my fairly locked-down home network. The amount of spew it continually issued onto the network was disgusting and I just disconnected it and drove in to complete whatever mandatory update they demanded. Instead of creating yet another goddamn "cybersecurity czar", how about we try to take what we've learned in the past 40-50 years in computer science and try to properly engineer the software foundations for the next 30 years? This is absolutely something the USG could instigate and accomplish, if the initiative was run by the right people. I know some of them!

by dsr_ 5y ago

snowwrestler 5y ago

OS does not matter anymore, it is a commodity. That’s why copies of any OS cost almost nothing at retail. What matters is the configuration of the OS and the network, and most importantly the selection and configuration of applications, whether hosted or client (or both). That is where DoD should focus their resources.

Jtsummers 5y ago

COTS is useful in many circumstances, it is much quicker to field if you can reuse existing commercial software. The only advantage to being incompatible with COTS is that it would feed the defense contractor industry as they'd need replacements or compatibility layers that some contractor would happily provide for $1 billion and 3 years (with an extra $100 million/year after it's late).

willcipriano 5y ago

> commission their own OS. I'd advise they start with a Linux distro. They can audit/replace components as they go to increase security over time. Even North Korea wasn't paranoid to build it's own OS[0]. [0]https://en.wikipedia.org/wiki/Red_Star_OS https://en.wikipedia.org/wiki/Red_Star_OS

jasonjayr 5y ago

Linux + wireguard-only network interfaces? With physically secured, uncloneable memory cards for each device's/interface's private key?

wyager 5y ago

> lots of jobs will be created Invoking the broken window fallacy is not a good sign for the coherence of your argument. I'm not entirely sure whether I should interpret this post literally or as a satire of NIH syndrome.

dsr_ 5y ago
Literal, please. Competing software ecosystems can constitute valuable economic activity. Broken Windows is, if anything, the current state of affairs: running important things on Microsoft Windows is becoming intolerable, with layers of expensive (dollars and computrons) fixes being applied. Is it economically better to start over or to keep trying to defend the indefensible?

41b696ef1113 5y ago

>Time for the DoD to commission their own OS. I am shocked that governments are not investing mega-bucks into getting a microkernel OS built that could be run on internet routers, tanks, power plants, aircraft, water treatment facilities etc. Even if microkerenel design has some impossible to overcome performance limitations, for utmost security, it would be a small price to pay knowing that hacking the OS was nearly impossible.

d_runs_far 5y ago
Your shock assumes there is leadership at the top that sees the big picture with respect to enabling technology. While not in the US, I have seen the inner workings of other nations armed forces, and there is an overwhelming urge to farm everything out to contractors - so there are going to be at least two competing microkernel designs (proprietary of course) for each the army, navy and air forces under the current leadership structures.
marcosdumay 5y ago
> Even if microkerenel design has some impossible to overcome performance limitations It doesn't. There's no inherent hit on microkernel performance. What exists are limitations of hardware that wasn't designed for it, and a complete lack of interest on creating hardware optimized for them.

V_Terranova_Jr 5y ago

My god, I cannot second this enough. We talk craptons about "innovation" and torrents of words on "cybersecurity" yet underpin daily life on recursive kludges on 1980s operating system and software technology. It's quite Kafkaesque! This isn't just a matter of "spending more on IT" - they likely already spend too much. It's just spending on the wrong things. No amount of layering crap on top will overcome the rotten, slapdash foundations of systems like windows, nor the diverging motivations between MS and and their customers. Especially customers like the U.S. Government. The amount of vendor lock-in to the shitty MS ecosystem, begetting atrocities like critical data living on godforsaken Sharepoint, is profoundly saddening. It's remarkable how everyone just uses hardware that's orders of magnitude faster than 1990s computers, yet interactions are far slower. I once tried (reluctantly) connecting my USG-issued laptop onto my fairly locked-down home network. The amount of spew it continually issued onto the network was disgusting and I just disconnected it and drove in to complete whatever mandatory update they demanded. Instead of creating yet another goddamn "cybersecurity czar", how about we try to take what we've learned in the past 40-50 years in computer science and try to properly engineer the software foundations for the next 30 years? This is absolutely something the USG could instigate and accomplish, if the initiative was run by the right people. I know some of them!