Hacker News new | past | comments | ask | show | jobs | submit login

Thanks, this is true. Encryption keys can exist in memory for up to a week [edit: or longer] in live memory with power on. This is well after they've been used on the machine, or say after you close your Truecrypt volumes. The vast majority of people rarely shutdown their laptops for extended periods of time and often just suspend to disk, instead of a full power-off.

This is why it's good to power down or sdmem when you're finished working with sensitive data.

On a full shutdown persistence is not as big of a risk, as the other commenter pointed out, cold boots are mitigated by DDR3 similar to how modern SSDs with TRIM make deleted data-recovery nearly impossible (such as Swap data which may also contain encryption keys).




> Encryption keys can exist in memory for up to a week in live memory with power on.

No. After some high-enough RAM area contained the keys and you keep it powered and your OS uses much less physical RAM than physically available there's no hard limit. Just forget the "week."


Updated comment for HN pedanticism.


You really should replace the words 'for up to a week' with 'indefinitely'. Mentioning a time limit at all is misleading in that context. This isn't pedantry: it's mere correctness.


Why would the encryption keys sit around in RAM? Properly designed cryptographic software wipes the keys when they are no longer needed.


In real life, keys actually can unintentionally remain in RAM even if in the ideal case they shouldn't. And you want more than just the keys to be wiped: your secrets you want to keep, the reason you decided to use Tails. And the fact that the Tails cares about implementing RAM wiping is the top post from dmix.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

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

Search: