Rich Freeman on 24 Oct 2017 14:06:43 -0700

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] Securely destroy and responsibly recycling hardware

On Tue, Oct 24, 2017 at 12:43 PM, Greg Helledy <> wrote:
> We've talked about that on this list in the past, the fact that your
> best-encrypted files of today could plausibly be easy to read 20 years from
> now.

I'm skeptical of that claim.  Certainly the best technology we have
today isn't intended to be defeatable in 20 years.  Nobody can predict
breakthroughs of course but it definitely won't be brute-forced in
that time.

20 years ago I think I was using 2048 bit RSA.  That certainly isn't
defeatable today, and even that will probably still be safe in 20
years, though I'd prefer a longer key length today.

I'm skeptical that 128-bit AES will ever be defeated.

> 20 years ago the dotcom era was getting underway and the internet really
> began to take off for the public at large

20 years ago PGP was old news, and SSL was common for commerce.
Neither is anywhere near being defeated today.  512-bit RSA is
certainly risky these days, but better options than that were PGP even
back then.

Sure, the average person on the street might not know anything about
it, but they hardly reflect best practices.  Most people still don't
know what AES is today.

And most of the weaknesses are in public key encryption.  I don't
think a serious symmetric cipher from 20 years ago is anywhere near
being broken (IDEA, 3DES, etc).  That is what you'd be using for
full-drive encryption, preferably with the key stored in a TPM (though
with the recent fiasco it might be better to not actually have the TPM
generate the key).

I'm not suggesting that people should be complacent, but throwing your
hands up because all modern crypto technology won't work in 20 years
is the wrong attitude also...

Philadelphia Linux Users Group         --
Announcements -
General Discussion  --