Where can I read about how Curiousity’s hardware and software works at a low level so things like communications still work in safe mode? Truly fascinating engineering.

thanks, ordered

Could someone who's read both compare the two books that have been linked? The !GP one sounds much more technical going by reviews on Amazon. The GP one sounds more about the organizational challenge (both are interesting).

I haven't read either, but Emily Lakdawalla's book is published by Springer Praxis, and I find that their books tend to be satisfyingly technical and detailed.

Not a read, but there was this cool conference called "The Mars Rover On-board Computer" last year at 35C3 :

https://media.ccc.de/v/35c3-9783-the_mars_rover_on-board_com...

Here is the source JPL article: https://www.jpl.nasa.gov/news/news.php?feature=7339

So how cautious are you when debugging a valuable system like this? I imagine you’re not allowed to just ssh and `sudo -u less logs`

'less' with 29 minute ping is no fun..

Try entering a password interactively /s

I wonder if they run a paralell earth based system to run simulations on first.

Yes - several, at varying degrees of fidelity, the highest of which is an exact replica rover named MAGGIE that lives in the garage next to the JPL “Mars Yard”.

TIL - that's really cool, though surprisingly difficult to find info about.

Debugging a system like this must very scary because of the tremendous responsibility. It is intersting to know how they actually communicate with it and handle auth.

It might not be authenticated. For some NASA probes, they just figured that any attacker would need control of a global network of high-gain radio antennas, and there aren't very many of those. Physical security at DSN tracking sites is sufficient in most cases.

I wonder if this threat vector will change as more countries become space bound. Curious to see the JPL Security Coding Guidelines, whenever that is written.

AES-256 on both uplink and downlink sides. It's fast enough and the NSA thinks it's good enough. Put a timestamp inside the cleartext to prevent replay attacks and re-key the encryption module every so often to prevent cryptoanalysis from breaking your current key.

Does anyone have more info on this? From what I recall there are mirrored systems with duplicate CPUs, memory, flash, etc. In the event of issues with one computer, they failover to the second. Was this an issue with something on the "A/primary", "B/backup" computer, or something more global?

Take a look at this : https://media.ccc.de/v/35c3-9783-the_mars_rover_on-board_com...

Around 25:45, there is a quick part about the CPU A flash issues on Sol 200.

Just restart the Docker container. Amateurs!

happy to hear comcast support was able to help out on mars. just unplug it and plug it back in.