Peter R’s Theory on the Collapse of Mt. Gox

Re:Peter R’s Theory on the Collapse of Mt. Gox by smoyer on Wednesday March 05, 2014 @06:41AM (#7346425)

Let's say I have a machine that's not connected to the Internet, but I develop a serial protocol that allows the transfer of bitcoins across RS-232, and build a cable that only has the receive signals connected at my "secure" machine.

If I wanted to take it further, the "secure" machine could print that wallet onto paper or could robotically insert flash drives into a USB port (in such a way that removed keys were dropped into a bin the robot couldn't reach into.

This is only one of many ways I can think of that would allow the automatic creation of cold wallets ... the only way to attack such a system is to gain physical access.

[ Parent | on HN ]

Re:Peter R’s Theory on the Collapse of Mt. Gox by Someone on Wednesday March 05, 2014 @02:59PM (#7349446)

You could still attack the stuff talking that RS232 protocol. Maybe it has a buffer overflow that allows me to gain control and change the software to write any money coming in not only to those flash drives but also out to my account (writing it to the flash drives makes it harder to detect this hack) In the limit, that doesn't require much. For example, I could flash a control LED and attempt to read it from across the street. Even if that succeeds in only a small fraction of transactions, it might still be worthwhile.

TL;dr version: anything that is connected to the outside world, no matter how small, is an attack target.

[ Parent | on HN ]