Hey all! New to the board here. I am looking to give a handful of talks on car hacking, Starting with an intro to car hacking and then progressing along the way. I currently use some basic arduino and cam-shield combos.
Anyone here have some advice on what a good place to start an intro talk is? Should I come all the way down to the bus/signaling level, or start higher up?
Thanks in advance!
-Z
The traditional approach is to go bottom up - start talking about the signals of CAN (briefly) then move up to CAN messages then build upon that to show what sort of things might be in messages, etc. But, to some extent car hackers probably care as much about how CAN looks on the wire as computer hackers care about how ethernet looks at a deep level. Which is to say, most of them don’t. The signals are unimportant, presumably the designer of the hardware knew what they were doing. Most people usually still mention signal level a bit to give some background but don’t spend a lot of time on it.
The really important aspect is the message format itself (std/etc, rtr, 0-8 data bytes, etc) then build on the abstractions above - J1939, CANOpen, UDS, etc. One needs to understand how UDS is formatted before getting into the nastier, more complicated aspects like challenge/response for security access.
So, I’d cover signals briefly, cover raw CAN for a little bit (it’s actually pretty simple and shouldn’t require much time I don’t think) then move to more and more complication and abstraction away from the raw format.
I think the specific conference matters a lot, but you probably don’t want to write a different talk for each one. So maybe write one long talk that spans from the basics to detailed examples. Then you can take the parts you want for each different conference depending on how much time you have and how much detail the audience wants.
I think it would be helpful to have a few demo videos showing traffic and you might want to go into how to pick out one signal over all the others so you can identify its function.