On Sat, 23 Feb 2002, Bill Stewart wrote: > If the ignition key crypto communications happen out at the steering > wheel, it's defeatable by basic hotwiring, but if they make the > communications happen from the electronic ignition module, that's > tougher to crack. The enterprising car thief _could_ carry around a set
Tougher indeed, if you decrypt the ignition table based on the secret in the car key transponder. Something like a nuke PAL. (They're not doing that yet, I know). > of EPROMs for different car models - or could resort to car-jacking, or They're not EPROMS. Last time I ran into them these are custom embeddeds (68HC11, MC68k derivates on a custom serial packet bus for BMW, nowadays they probably will use ARM), with most code in PROM, some EEPROM (nowadays flash, I guess) and some RAM. You'd need an entire part, and they're not available on the open market. (Assuming, you'll go to the pains of driving up a flatbed truck to the parking lot, and a specialist who can exchange and program controlled parts, which really asks for high end cars for the effort to pay). > social-engineering at parking lots. Fancy electronics don't know that Looks easier that way. > you stole the keys. > > But those attacks are more trouble than stealing an unattended car, > and work equally well against non-cryptographic cars, > so it's a real risk reduction.
