Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Not really, if this attack is doable over the internet, which will have latencies randomly[1] distributed (in a normal curve), rand does not actually do anything, well maybe require a few more sample points to get that noise out.


That makes sense, maybe sleep(rand(1.0 - time_spent))


Or just sleep(1.0 - time_spent)


Or just don't short circuit ever, and do the entire computation every time.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: