- cross-posted to:
- space@lemmy.world
- hackernews@derp.foo
- cross-posted to:
- space@lemmy.world
- hackernews@derp.foo
You must log in or register to comment.
I was listening to the Security Now podcast a couple of months ago and Steve Gibson spent a fair amount of time talking about the challenges arising from trying to communicate with a spacecraft from that far away. Random but fips can happen from all the crazy space radiation along the way or it can just be written to memory incorrectly because of said space radiation and you’re not going to replace the hard drive on a 1977 space computer that far away from Earth.
Last time they had an issue with data corruption the communication antenna ended up pointed the wrong way and they had to “shout” the command and hope that the probe would “hear” it so that they could get the antenna pointed back the right direction.
Hard to imagine that those 2 have been out exploring space for 46 years.
Random bit flips are fairly common on Earth, too. This is why ECC memory should be standard equipment on systems handling data that anyone cares about, rather than reserved for servers.
It can happen at any stage where transistors are involved. So for important thing just ecc memory is not enough, I saw a video where newer space craft basically have a voting system, so 3 or 4 subsystem calculating exact same instructions, and the “majority” wins.
Fun fact! We actually do this voting system with nuclear power PLC’s in my industry. This sort of redundancy is used in all sorts of safety controllers
Aaaaw 🥹