Re:Timing Attacks Explained
by fragmede on Tuesday July 27, 2010 @12:32PM
(#1551745)
>You're right of course. So if it was done in say multiples of 15msec. The first window for a response is at 15msec. The second window at 30msec. The third window at 45msec and so on.
Instead of the comparison that checks the whole field, start two threads, one per real cpu core, and run the unknown comparison against an a known good, and in parallel, and only return when both have completed. Not the most efficient, but would scale properly for faster and slower cpus.
Instead of the comparison that checks the whole field, start two threads, one per real cpu core, and run the unknown comparison against an a known good, and in parallel, and only return when both have completed. Not the most efficient, but would scale properly for faster and slower cpus.