3 ms·

> Yes, someone could jump in and do something bad, but the greater danger is the encryption/lock system causing an accident by getting in the way of legitimate

> Yes, someone could jump in and do something bad, but the greater danger is the encryption/lock system causing an accident by getting in the way of legitimate commands. That's just not how encryption works... Also, a malicious actor being able to point a multi-ton rocket filled with explosive fuel the wrong direction would be far worse than the system continuing the mission autonomously. Now imagine human spaceflight missions where the command stream can be hijacked by anyone with an SDR. Yeah, not happening. The command stream better be authenticated, and if you're doing cryptographic message authentication, you might as well encrypt the payload so no one else can read it anyways. > The rocket might be designed to ignore commands coming from the wrong direction and/or signal strength. You want to talk about something error prone "getting in the way of legitimate commands"? That would be an unbelievably error prone approach. Oh, the rocket accidentally rolled on its axis a few degrees? Ignore all commands to correct it! Encryption and authentication are extremely reliable.

sandworm101 6y ago
>> spaceflight missions where the command stream can be hijacked by anyone with an SDR. Yeah, not happening. An SDR attached to some serious antenna hardware, enough to get into the sidelobe of an antenna pointed at the legitimate command network. As for rockets rolling, that isn't handled by ground commands. That sort of stuff it totally within the rocket's internal systems. The ground staff have very little control until the second stage has done its thing.
zelon88 6y ago
> Now imagine human spaceflight missions where the command stream can be hijacked by anyone with an SDR. Yeah, not happening. I am no expert on spacecraft command channels but I would highly doubt that a lot of non military US spacecraft use encrypted command signals until extremely recently. He is right. Space engineers are extremely adverse to trying new technology and technique. They are even more adverse to adding code to things that work without adding more code. In order to talk to one of these things you have to have one of the satellites in the NASA DTN. Technically the DTN protocol supports authentication and encryption but remember that everything about DTN must maintain perfect backwards compatibility with space assets that are made with decades old technology. Security is not it's primary goal like with terrestrial communications. Connectivity is far more important.
- sandworm101 6y ago
  Some US hardware also operates on frequencies that do not penetrate the atmosphere, an old failsafe against ground-based interference/spying. So, even if totally unencrypted, the attackers SDR would have to be in orbit. That's a high barrier for the average hacker.
  - beerandt 6y ago
    Well yeah, but the actors that can overcome that barrier are likely the ones you're most worried about in the first place. So maybe it limits number of adverse actors that can attempt anything, but it doesn't allow you any shortcuts in system design/security. I've always assumed that frequency barrier stuff was more about defense against tactical, air based electronic jamming/countermeasures. ETA: Thinking about it some more, what you describe seems like a slightly different type of security by obscurity. If you're relying on it for security.
    - sandworm101 6y ago
      Obscurity is about hiding amongst the noise. I describe physical network separation, essentially air-gapping. Do we encrypt the signal between a car's brake pedal and the brakes? Why not? It is a very important network that could result in loss of life. We don't encrypt because the network is physically separate from attackers. If they are tapping into your car's internal wiring then they could kill you in any number of easier ways. A rocket tuned to only listen to emitters from specific locations need not encrypt because it is also physically separated from potential attackers. So in high-reliability systems if encryption/authentication only offers protection in hypothetical edge cases where the attackers is co-located with controllers, but presents any additional complications, it won't be employed.
      - perl4ever 6y ago
        >We don't encrypt because the network is physically separate from attackers. Given the existence in modern vehicles of: - a shared bus for communication amongst the processing units - autonomous emergency braking being fairly standard and - network connectivity being available in some cases I think your analogy might be outdated.

sandworm101 6y ago

>> spaceflight missions where the command stream can be hijacked by anyone with an SDR. Yeah, not happening. An SDR attached to some serious antenna hardware, enough to get into the sidelobe of an antenna pointed at the legitimate command network. As for rockets rolling, that isn't handled by ground commands. That sort of stuff it totally within the rocket's internal systems. The ground staff have very little control until the second stage has done its thing.

zelon88 6y ago

> Now imagine human spaceflight missions where the command stream can be hijacked by anyone with an SDR. Yeah, not happening. I am no expert on spacecraft command channels but I would highly doubt that a lot of non military US spacecraft use encrypted command signals until extremely recently. He is right. Space engineers are extremely adverse to trying new technology and technique. They are even more adverse to adding code to things that work without adding more code. In order to talk to one of these things you have to have one of the satellites in the NASA DTN. Technically the DTN protocol supports authentication and encryption but remember that everything about DTN must maintain perfect backwards compatibility with space assets that are made with decades old technology. Security is not it's primary goal like with terrestrial communications. Connectivity is far more important.

sandworm101 6y ago
Some US hardware also operates on frequencies that do not penetrate the atmosphere, an old failsafe against ground-based interference/spying. So, even if totally unencrypted, the attackers SDR would have to be in orbit. That's a high barrier for the average hacker.
- beerandt 6y ago
  Well yeah, but the actors that can overcome that barrier are likely the ones you're most worried about in the first place. So maybe it limits number of adverse actors that can attempt anything, but it doesn't allow you any shortcuts in system design/security. I've always assumed that frequency barrier stuff was more about defense against tactical, air based electronic jamming/countermeasures. ETA: Thinking about it some more, what you describe seems like a slightly different type of security by obscurity. If you're relying on it for security.
  - sandworm101 6y ago
    Obscurity is about hiding amongst the noise. I describe physical network separation, essentially air-gapping. Do we encrypt the signal between a car's brake pedal and the brakes? Why not? It is a very important network that could result in loss of life. We don't encrypt because the network is physically separate from attackers. If they are tapping into your car's internal wiring then they could kill you in any number of easier ways. A rocket tuned to only listen to emitters from specific locations need not encrypt because it is also physically separated from potential attackers. So in high-reliability systems if encryption/authentication only offers protection in hypothetical edge cases where the attackers is co-located with controllers, but presents any additional complications, it won't be employed.
    - perl4ever 6y ago
      >We don't encrypt because the network is physically separate from attackers. Given the existence in modern vehicles of: - a shared bus for communication amongst the processing units - autonomous emergency braking being fairly standard and - network connectivity being available in some cases I think your analogy might be outdated.